Growth forecasts and government
I was just reading a post on forecasts for US economic growth at Econbrowser. In it the author says: “However, even back in December, the White House forecast was slightly more optimistic than the private sector consensus”. This ‘overconfidence’ in the economy seems to have been a common theme in US public sector forecasts over recent years.
Compare this to Treasury forecasts of the New Zealand economy, which have been consistently below the private sector consensus – that is the reason why tax revenues have consistently surprised on the upside in New Zealand.
Here we have two government authorities, one which constantly overstates economic growth and one which understates economic growth. Why do you think we have this difference?
Update: My brief thoughts are below the flap:
I’m putting it down to one of two things:
- A bias associated with the economists belief of what the ‘mean’ is. In this case the economist over-reacts to short-run information when setting the ‘mean’ value for the model to revert too in the long run.
- Economists inability to recognise ‘structural shifts’ (to be fair, who the hell can observe structural shifts every time!). A structural shift can move the ‘mean’ value of some economic variable. If economists fail to recognise this, then they will will have the model reverting to the wrong point.
I’d check your second statement, I don’t think Treasury has been consistently below the private sector consensus, the consensus was just consistently too pessimistic (and even then, I think a lot of the surprise has been because revenues have generally been stronger than nominal GDP would historically imply).
As to the first, I don’t know what the long-term performance of the government forecasts are like, it is supposedly an independent body I believe.
I suspect that most forecasting models, unless they’re reviewed constantly, will have a tendency to produce periods of forecasts where the errors all in the same direction. Alternatively, most models to have a lot of subjective input, and that input might be persistently too optimistic or pessimistic.
“I’d check your second statement, I don’t think Treasury has been consistently below the private sector consensus, the consensus was just consistently too pessimistic’
Hi, I realise that the consensus was too pessimistic, however I thought we said something in our last forecast round about Treasury forecasting consistently below the consensus – or maybe that was just consistently below our forecasts 🙂
“I suspect that most forecasting models, unless they’re reviewed constantly, will have a tendency to produce periods of forecasts where the errors all in the same direction.”
I think your reasoning is right, after all most forecasting relies on some form of mean reversion, and when the mean is only occasionally revisited we are bound to end up with some form of consistent bias between times.
“As to the first, I don’t know what the long-term performance of the government forecasts are like, it is supposedly an independent body I believe.”
True, but my interest lies with what different factors have caused this implicit bias in forecasts. Both the US and NZ have huge current account deficits, and experienced a significant housing bubble over this time – what underlying assumptions have lead to one body over-estimating growth while the other under-estimates growth?
I think the answer is that the U.S. FEd and white house lackys etc are still trying to fool the public that they are not tettering over a cliff, the height of which is not really known, and they as in all things, want to be the controlling factor “when”; so it can be decided and co-ordinated with other going ons when the fall happens.
Back here in New Zealand, the under-estimating is more about the govt. trying to fool itself that it’s approach to economics has been coherent and remains beneficial to the society, i.e. they understand what they are doing…
meanwhile back on the farm, due to the consequences of the most elementary flaw of our Debt based system not being recognized and addressed, the ratio of wolves to sheep steadily increases day by day in a rich land.
DSC 08.