Craven’s Wager

My friend John Bevan sent me a link to this video by Greg Craven, and asked me to comment on it, so here you go.

In the video, which has been around for several years and has several million views (by which I mean to point out that this contribution is superfluous), Craven gives his analysis of the climate change debate. Using a simple logical tool, he seems to cut straight through all the controversy over whether and to what extent climate change is man-made and preventable, to an incontrovertible conclusion about what we must do: act now to prevent it.

The logical tool Craven uses is a 2 x 2 table to compare the four possible worlds arising from the questions, “is climate change man-made and preventable?” and “will we act to prevent it?”:

Craven’s argument is that debating whether climate change is man-made and preventable, or even debating whether there’s a genuine controversy, is worrying about which row we’re on. Even in the absence of a conclusion about that, we can still decide which column we want to be on. Do we want to choose the outcome which gives us ‘Cost and (economic catastrophe or civilisation saved)’, or the ticket which gives us ‘civilisation continues or massive civilisation-destroying catastrophe’?

At the moment, inertia over the row debate means we’re choosing the second column by default of inaction, but when you look at it this way, the only reasonable decision is to choose the first column – act against climate change even if we’re not sure it’s happening – because the overall wager is better.

John agrees with the conclusion, that we should act against climate change, though he was uncomfortable with the argument, worrying that it sounded too much like Pascal’s wager.

Pascal’s wager is the argument in favour of theism, which proposes a similar tabular analysis of whether god exists, and the consequences of believing in him:

There are a number of problems with Pascal’s wager, but I don’t think they apply to Craven’s wager.

  1. The action required to choose the first column is “believe in god”. You can’t put yourself in it just by going through the rituals of prayer, church attendance, and so on. It’s difficult to see how you could honestly ‘choose to believe’ based on an assessment of outcomes, rather than evidence.
  2. It overlooks the fact that there are numerous gods to choose from, most of whom demand exclusive belief in and worship of them in order to get their reward. In other words, it’s not a 2 x 2 table, so the choice of the belief column isn’t obvious.

Now, instead of a straight-forward choice of ‘eternal salvation or time wasted’ over ‘time well spent or eternal damnation’, it seems as if every ticket is probably going to damn us, and the tool becomes useless for decision-making.

Clearly, the first objection doesn’t apply to Craven’s wager. By choosing the first column, and acting against climate change, we’re not required to choose to believe with absolute certainty that climate change is happening: only to make certain concrete and achievable policy changes.

I agree with John that Pascal’s wager is a flawed argument, but that doesn’t mean that wager arguments in general are flawed. Pascal’s use of one is flawed because of the choice/belief problem at the very least, but in other contexts without that problem, the wager analysis is sound, and we use it all the time. Our decision to buy insurance policies, for example, is an intuitive use of a wager analysis, and I think Craven’s wager is analogous.

We don’t know whether we’ll have a catastrophic accident which will cost us thousands of pounds, but either way, the better choice is to buy insurance: to choose the first column, where we will definitely lose £100 but at most £200, rather than risk losing thousands. Craven is arguing that we should act against climate change not because we know we have to, but because it’s a good insurance policy: it’s better to pay the relatively limited cost of economic recession to implement the recommended changes, because it guarantees us against the catastrophic results of unmitigated climate change.

What about the second objection? Craven’s argument doesn’t take into account the law of unintended consequences. The recommended actions to take against climate change are based on the same science which claims that climate change is happening. If we take those (choose the first column) and it turns out that science is wrong (we’re in the first row), then the results may be even worse than simply the economic effects of the policies. We may bring upon ourselves climatic changes which wouldn’t have happened otherwise, so the worst results of that ticket are as bad as the alternative. So, as in the expanded Pascal’s wager, should Craven’s wager have numerous columns representing different actions we can take, and numerous columns covering various catastrophic consequences of them going wrong?

It looks like Craven’s wager is now as useless as a decision-making tool as the expanded Pascal’s wager. Every choice has unknown consequences ranging from massive catastrophe to no effect.

However, what are the possible policy choices being proposed, and are their consequences really as unknown and potentially catastrophic as we’re supposing here? Craven isn’t using his wager to argue for anything as drastic as climate engineering with stratospheric sulfate aerosols, which have the potential to negatively impact the climate in ways we can’t predict. He’s only arguing for a combination of scientifically uncontroversial measures, like reducing our energy usage and converting to less polluting forms of energy production. Since the effect of such measures is only to reduce our impact on the planet to its earlier low levels, then the unknown consequences across the top row (ie, of taking action when none is required) do not really range into the catastrophic. Further, he’s arguing that we take all relatively uncontroversial measures (ie, no to experimental climate engineering, but yes to more renewable energy, and industrial regulation, and carbon capture, etc, etc), not just some, so we can re-merge all those columns and rows.

The problem of unintended consequences remains, but what is the best way to deal with that? Luckily, unlike Pascal’s wager, where the consequences of believing in the wrong god are only revealed when you die and it’s too late to change your choice, action on climate change will have consequences in the real world, which can be observed, and our policies and actions can be changed in response. It’s not a one-time gamble between the different policies. We can – and according to Craven’s wager should – implement them all, and continue to investigate their effects, testing and adjusting over time. The only sensible choice is to base our policies on an ongoing, comprehensive and rigorous programme of observation: science.

One thought on “Craven’s Wager

  1. The issue I have is that in the above you’ve said that this makes sense once you factor in knowledge / probability of the various outcomes. The video stated that you should ignore knowledge/evidence and just go for the square with the happy face, despite there being a better square available through taking no action. You don’t need the grid to determine the course of action, you need the facts – you can use these on their own or with the grid to make your decision, but without them the grid alone, to me at least, implies the best option is to take no action and hope that things work out; any other option leads to some form of catastrophe, albeit one of which includes a smiley.

Leave a Reply

Your email address will not be published. Required fields are marked *