I've seen odd claims about "rules vs. discretion" a couple places now, so I thought I might comment on this. First, on a facebook, a relatively well known economist that everyone here would know recently equated "discretionary policy" and Keynesian policy in reference to Kydland and Prescott (1977). Then just yesterday Don Boudreaux acted as if Nicholas Wapshott, in suggesting that Milton Friedman would have advocated greater steps to "manage the economy" (Wapshott's words), was saying that Friedman would have supported discretionary policy. Don's response was that Friedman was an enemy of discretion and preferred policy rules.
Don, I think, is trying to pull the wool over his readers eyes here. The consensus for a long time now has been that policy rules are superior to policy discretion. All the major Keynesian voices Don usually complains about, and those who he would associate with Nicholas Wapshott, support policy rules - not discretion. Paul Krugman has specifically cited a Taylor rule to govern monetary policy, and has worked out a Mankiw rule answer in the past as well. Brad DeLong has come out in favor of an NGDP level target, although he's made points in the past grounded in the Taylor rule. And then of course there's Mankiw who is presumably pro-Mankiw rule and John Taylor who is presumably pro Taylor rule.
In short, Keynesian policy is rules-based policy. People might just cite the need for stimulus right now, because the shortfall is so egregious. But when you ask Keynesians what is required, a rule of one form or another is invoked.
Now there are important discussions going on about what rule is best, and level rules of any sort (NGDP level, price level, etc.) have important properties that rate-based rules don't have, particularly when we get knocked off trend so severely. It's important to have these discussions. But nobody is advocating discretionary policy.
To be honest, I doubt anyone ever did support discretionary policy. It's true, before Barro-Gordon and Kydland and Prescott the rules people followed were probably more ad hoc. It was probably more of the form "estimate the output gap with this proceedure and then fill the output gap". And since economic science and engineering get better over time, there are very real ways in which our rules are better than those old ad hoc rules. But I doubt anyone ever really advocated discretionary policy in the way that it's sometimes presented. The thing is, nobody had a reason to talk about "rules vs. discretion" until the rational expectations revolution made "rules vs. discretion" a thing. And then to make their case, they labeled the old guys "discretionary" and the new guys "rule based".
And that case was an important case to make, and the rhetoric used was effective for making it.
But I doubt there was ever a such a really stark divide on the question. People were always more or less rule based they had just never thought about the issue in those terms before.
4 hours ago