Richard Murphy says there is a "near inevitability" of a recession sometime soon. I'm not sure it's wise to make such a prediction.

I wholly agree with him - and with the IMF - that there are reasons to be worried. In fact, I'd add a couple to Richard's menu:

- The US breakeven inflation rate, at 1.2pp, is close to a six-year low. It implies that markets expect the Fed to undershoot its inflation target - which means bond markets think monetary policy will be too tight.

- At low inflation, the yield curve might no longer be the great predictor of recessions it has been in the past. The fact that the yield curve is upward sloping is therefore not as good a comfort as it once was.

However, recessions are damned hard to foresee. Most professional forecasters - even before 2008 - have failed to see them coming. This might be because they are inherently unpredictable because they arise in part from failures of individual companies which might be magnified by network (pdf) effects. Because we don't know enough about these complex processes, time-dated forecasts are, as David says, "a mug's game."

Yes, a recession is possible. But there's a good reason to be wary of forecasting one now. Share prices tend to do much better from Halloween to May Day than in the other six months of the year. This suggests that people's assessments of economic prospects usually become too gloomy at this time of year. We should, therefore, be especially sceptical of pessimism now.

Of course, if he keeps forecasting a recession, Richard is certain to be right eventually. As Andy Haldane said last month:

Recessions occur roughly every 3 to 10 years. Over the course of a decade, they are overwhelmingly more likely than not.

But this is like a clock that's right twice a day: it isn't good enough for practical purposes.

My big problem with Richard's forecast, though, isn't that it might be wrong: it might not be. Instead, I've two other beefs.

One is that forecasting a recession gives the many non-economists who read Richard a false impression that economists can foresee the future. In fact, we can't. Knowledge of the future is a contradiction in terms.

Secondly, I'm not sure Richard has got the cost-benefit calculus right.

His main ideas - some of which such as opposition to austerity in its current form, a state investment bank and the possible need for alternative monetary policies are good - don't derive their strength from the claim that we're heading for recession. Merely warning of the possibility of recession is good enough. However, if we don't get a recession, his credibility will be called into question to discredit these ideas - whilst if we do get a recession his call will be forgotten or dismissed (rightly) as a lucky guess. He's therefore giving his opponents a stick to beat him with, and unnecessarily so.