What is policy? Or how design is already showing its policy-making benefits
‘Policy’ is an easy thing to talk about and should be easy to define. We use it to discuss what the government is doing, decisions by our bosses at work, and even — jokingly — how we deal with things day-to-day. But we use it in such broad ways that it can be difficult to identify what we’re really talking about.
One of the benefits of using design in public administration has been in asking ‘What is policy?’ Designers working on policy have written about it, and the Policy Lab has delineated policies by type. It’s this ‘What tools are we really using and how do they affect what we’re actually doing?’ attitude that I really like about using design in public life. The world is so complex that I at least want to be clear on the concepts we’re using.
It’s a simple question, but think about what might be involved in a policy and it quickly gets complicated. Is something a policy if I only do it once? Should I only do it in one set of circumstances and what should happen if they change? Most importantly, how is a policy different from an act, plan or strategy?
The difference between an act and a policy is the easiest to define. The former happens only once. If on a whim I give money to one homeless person and no more, that’s not a policy. The conditions prevailing when I gave the money will repeat themselves many times over, but if I don’t react in the same way each time my behaviour can be said to be no more than impulsive. When Donald Trump fired missiles at Syria without explanation, the commentary afterwards asked if he would do it again and hence whether it was more than a one-time event — we wanted to know whether he had a policy to launch an attack every time Bashar al Assad used chemical weapons.
So that’s the first part of the puzzle: a policy has to be something that’s done more than once.
A policy is also not a plan. The point that military plans never survive contact with the enemy is true because plans are designed to work in only one set of conditions, and often ones over which there is a high degree of control. When I mention that I’m going to walk to the pub on Sunday afternoon, I say ‘I’m planning to’ because I know that the conditions are reasonably stable. The pub will be there, I won’t be working on Sunday, and I control my behaviour. If it rains and I don’t go, I say ‘I changed my plans’. A policy in this context would only make sense if I were to go walking in different places every Sunday and I said I would go to the first pub I encountered. In some places there won’t be a pub, and it would be absurd for me to say that I’m planning not to go to the pub when I regularly do.
So that’s the second part of the puzzle: a policy is designed for more than one situation, and will have a view on what to do in at least two of them.
But what’s the difference between policy and strategy? They both operate more than once and across more than one set of conditions, and of course with an aim in mind. Game theory is the best way to separate the two.
Game theory gives us a way to think about behaviour in a situation with more than one period, more than one player, and where the actions of other players change the rewards available to me in what I choose to do. Crucially, in order to have a strategy for whichever game I’m playing, the decision rule within it — ‘When X happens, do Y’ — needs to determine my actions in all circumstances. If I’m in a competitive game and have no response to something that happens, I’ll get punished and all of my previous effort will go to waste.
When the term is properly used, this is what strategy means. We say that we have a ‘corporate strategy’ or ‘industrial strategy’, meaning something that guides everything we do and at all times. If we had an industrial strategy for all sectors apart from, say, satellite manufacturing, we would change the description to show that we’re not accounting for that industry.
And that gets us to the definition of a policy: something that we intend to do with an aim in mind, more than once and across multiple periods, but not in all situations.