Infometrics chief forecaster Gareth Kiernan (File photo). Photo: Supplied
Every so often you see a headline: A major bank has trimmed its forecast for the official cash rate (OCR). An economist thinks the inflation rate will be higher this quarter.
But while the predictions make good news stories, what is the science behind them? And do they matter?
To find out, RNZ asked Infometrics chief forecaster Gareth Kiernan for his forecasting strategy.
He said each thing that an economist might forecast had its own nuances, so for the purposes of this exercise, we are looking at forecasting the OCR.
To come up with a view on where it might go from here, Kiernan would look at a few things.
That would start with Infometrics' pre-existing view on GDP - the size and growth of the economy - and the rate of inflation.
These are two important factors for the Reserve Bank when it sets the OCR.
The OCR is its tool for keeping inflation in check, and if GDP is increasing, it is likely to mean higher inflation.
Kiernan said a useful starting point was the Taylor Rule, from an American economist in the 1990s, which proposes that interest rates should be set based on how far inflation is from the central bank's target, and how far economic activity is from a sustainable level.
The Reserve Bank's target for inflation is 1 percent to 3 percent.
"This rule provides a useful baseline for where interest rates are likely to head over the next 12 to 24 months. Another factor that influences our modelling are estimates of what the neutral interest rate is, although that tends to be more of a factor in medium-term forecasts rather than short-term forecasts."
The neutral OCR is the level at which it is not putting pressure either way on the economy - not encouraging or discouraging consumption.
At the moment, it is thought that neutral is about 3 percent.
"With the OCR, there's an element of circularity in the forecast process, because higher interest rates in response to higher inflation and/or faster economic growth will naturally dampen future inflation or growth outcomes. Modelling can take this into account and converge towards an equilibrium outcome," Kiernan said.
"In any forecast, there is also a degree of judgment involved, and with the OCR, that level of judgment is probably higher than for many other economic variables."
That was because while forecasts for things like retail sales involved the actions of a large group of people who might tend towards some "normal" average, the OCR was set by the Reserve Bank monetary policy committee.
"We're forecasting the decisions of a small group of individuals, which potentially provides more scope for unusual behaviour.
"Thus there's an important element of combing through the Reserve Bank's statements and forecasts to determine the most likely path of action, particularly over the next 12 months.
"This judgment can mean that, although the modelling suggests the OCR should only be cut to 3 percent, there might be sufficient signals in the bank's communication that lead us to forecast further cuts to 2.5 percent instead."
He said economists' judgement would come into play more broadly because the outputs from the modelling process were only as good as the assumptions behind the inputs and the model itself.
"During times of unusual volatility or behaviour, models are unable to fully adjust to likely outcomes.
"At other times, we might think that the pass-through of changes in input variables might be smaller or bigger than normal and adjust appropriately.
"This is where experience of previous cycles - even though no two cycles are the same - and the ability to brainstorm and bounce ideas off others in the team can be valuable.
"Ultimately, from a pragmatic point of view, as a forecaster and presenter, I need to be able to stand in front of an audience and be confident in the story I'm telling them.
"Numerical model outputs are a valuable starting point and can be useful for challenging your thinking and pre-existing ideas and biases. But the reality is that the numbers are almost always going to be wrong, to some degree.
"Providing people with the story and thinking behind those numbers is equally, if not more, important, as well as highlighting issues that might change the outlook and mean the forecast numbers end up being wrong. Those issues and risks enable people to better respond to the changing environment, rather than simply taking our numbers as gospel - or, alternatively, simply rejecting them because economists just make everything up."
So, do forecasts have an influence on what actually happens?
Forecasts can have an impact on wholesale market pricing, particularly if there is strong consensus.
At one point last year, the wholesale markets had moved to such an extent that some commentators suggested the Reserve Bank had to move in the direction markets were expecting, to avoid unwieldy disruption.
Economist Shamubeel Eaqub said that was probably overstated.
"If the RBNZ disagrees then it will act accordingly. But if it is a matter of timing, they may choose to cut/hike faster to meet market pricing so the 'retail' signals don't work against them. But I think that is more tactical than being forced by the market per se."
Sign up for Ngā Pitopito Kōrero, a daily newsletter curated by our editors and delivered straight to your inbox every weekday.
