War is a very particular accounting issue

For me, like Chris, [going to war] is an accounting issue.

Thus sprach Paul. But I think that’s an ugly, confusing way of phrasing it. In retrospect perhaps you can decide whether a war was a good idea with a quick tally ex-ante and ex-post death tolls with some role for economic performance, but a forward looking evaluation is very different.

My priors suggest strongly that killing people is very wrong. My priors are very strongly biased in that direction. Call it a personality flaw! I think you’d have a tough job convincing me that killing one person was worth probably saving a hundred.

Thinking about war as something likely to be fucking terrible and which definitely has incredibly high costs and definitely has incredibly uncertain benefits is useful. It sets the bar for war incredibly high. No particular policy position needs be behind the war or your opposition, I am just talking about stacking the deck.

Stacking the deck gets a bad name, but stacking the deck against killing people is a self-evidently pretty good idea. War is one subject where I think the principle “first, do no harm” is a good one.

Why? Because history. The historical record says that some truly awful actions can lead to prosperity but that truly awful wars normally just lead to death and suffering. So, for example, my priors tell me to somewhat support what’s happening in coastal China, but not what happened in Iraq. I’ll do the same in the future too.

Do no harm? You have no choice


The principle “first, do no harm” is a good one in most situations. But one situation where I find it hard to apply is in development: getting poor countries rich. Harm is inevitable and unavoidable and to some degree unmitigatable. Owen Barder lays it out thus:

We should pursue policies which do (much) more good than harm, which take account of the broader consequences, which take proportionate steps to limit the harm and risks, and which are careful and responsible about looking after those who are likely to lose from change.

I think is an eminently sensible position to take. I’m going to make a more fundamental argument that Owen though. Sometimes, merely moving policy in the right direction will have directly harmful side effects which you won’t be able to mitigate and this is okay. It’s okay because we don’t really have another option.

Norm takes umbrage with Owen’s defence of a utilitarian approach to the harm caused during development (and will likely dislike mine), laying out two key reasons. First, there are some things which are so wrong that no amount of mitigation can offset, Norm offers the example of torture. The other is that far-off, uncertain, broadly shared benefits should not be won at the cost of concentrated, short-lived, acute, concentrated costs.

Neither of these are good arguments for adopting “first, do no harm” in the field of development.

Take something as innocuous sounding as quinoa (no, really, take it!). According to the Guardian the trendy vegan boom in quinoa consumption has meant that the poor Bolivia and Peru can no longer afford their staple food and must seek alternatives. Although the guardian gives the impression that harm is occurring here, it is important to remember that we are talking about a mitigatable harm. Redistribution within Bolivia could split the gains from trade between winners and losers, with everyone better off.

But think more deeply about what’s going on here. Neither Bolivia nor Peru has reached the technological frontier for quinoa growing. I’m only working on the photos of those growing it supplied in the Guardian, but it looks pretty low productivity to me. Prices have increased because it is difficult to increase production to keep up with demand. Development is the process of ensuring that production eventually does.

However, this poses a paradox. Actually existing institutions that have seen this happen have also created incentives to do great harm.

Meeting demand (for export markets and domestically) will require investment. Those doing the investment need a degree of certainty that those investments will get a return and this means securing property rights for investors so they can spend big now and reap the rewards long term.

But, if the institutions protecting the property of the wealthy are stronger than those protecting the property of the poor then those peasant cultivators of quinoa are going to be kicked off their land. Given the institutional strength of the wealthy in any polities and the compromises unavoidable in politics the decks are likely stacked against those farmers. This is a harm that could be, but will not be mitigated.

I don’t just refer to simply exploiting workers, but to the expropriation of land and the intimidation of communities. These things are typical of development success stories because effective institutions make the stakes so high. They make honest hard work pay, but they often make the criminal exploitation pay too. Think Chinese land grabs, or America’s westward crawl – the historical record is full of successful developmental cases where incentives to profit coincide with incentives to pillage.

The job of development is to apply pressure and create institutions that cause economic growth, but brutality and suffering has coincided with all stories of development. “Do no harm” doesn’t work where successful policies cause suffering incidentally but inevitably.

Do I like it? Not exactly. Do I have a choice? Again, not really. Is “first, do no harm” a good principle for guiding development policy? Nope. I wish it were, but all the maps I have from getting from rural idiocy to modern society have these satanic mills halfway. [1]


[1] Actually, there is a detour called “migration”, but that’s another blogpost for another time. In brief, if the institutions won’t come to Mohammed, Mohammed should go to the institutions.

[2] I find the this excerpt (below) at odds with Norm’s support for certain Mesopotamiam adventures as they featured a likely high ex-ante and a realised high ex-post death toll, but that is a subject for yet another blog post.

There are doubtless more qualifications necessary than just these two, but the point is that the ‘Do no harm principle’ can’t be adequately replaced by a purely quantitative measure of goods against harms. Some harms should not be done in any circumstances  Some benefits are too uncertain to justify definite harms.

Postscript: I have missed blogging. Ah. Feels like I’m stretching after a long lie down.