Westpac predicts employment will remain strong: Justin Smirk

EXPERT OBSERVATION

Unemployment fell to a six year low of 5% in September. If you look at the labour market holistically, the large monthly drop looks like survey volatility rather than a new lower trend. Nevertheless, it is our view that the labour market is sound and that unemployment has been trending down.

Westpac forecasts the unemployment rate to hold above 5% to end 2018 before a soft patch in the June quarter 2019 (due to both election uncertainties and correction to the 2017/2018 strong patch) will lift unemployment to 5.3%. Once past the soft patch, steady employment growth will take unemployment back down to 5.0% by end 2019 and 4.8% by end 2020. 

We use business and household surveys to guide our near term employment forecasts while our medium term profile is driven by our fundamental view on household demand, services exports and non-mining investment.

Our participation forecasts are dynamic responding to changes in employment growth. The forecasts are, however, bound by the assumption that participation will not hit a new record high.

Household and business surveys guide the near term

When forecasting the near term pace of employment growth we use the labour market indicators from the household and business sentiment surveys.

From the Westpac-Melbourne Institute Consumer Sentiment Survey there is the question on household unemployment expectations (do they expect unemployment to rise or fall?) which has a good fit with the momentum in the labour market. Unemployment expectations have risen modestly (households expect unemployment to rise a little thus a softening in the labour market) but still holding a historically robust level. However, employment overshot unemployment expectations through the second half of 2017 so an undershoot through to the mid 2019 would be a normal course of events.

The multiple business surveys in Australia have questions that are applicable to the labour market. To generate a broad, deep labour market indicator, Westpac compiles all the relevant indicators from these surveys into the proprietary Westpac Jobs Index from which we generated a simple model of annual employment growth. It is true that the Jobs Index is more stable than the annual pace of employment growth but we have found it provides a useful guide on both the near term pace of employment growth and, more importantly, major turning points. Currently the Jobs Index is pointing to employment growth of 2½%yr. Actual employment growth in the second half of 2017 overshot the jobs index. Looking ahead we expect employment growth to undershoot the job index through to early 2019.

Using these indicators we expect the pace of employment growth to hold around current levels through to the end of the first quarter of 2019. This will see the annual pace dip to around 2% by end 2018 which is softer than what our indicators are suggesting in response to our allowance for some employment undershoot following the 2017 overshoot.

Looking further out household demand and non-mining investment drive our forecasts

Participation has peaked, for now, but will recover as employment recovers

Once you have an employment profile, an estimate of participation is required to round out the forecasts. As readers would be aware, the strong employment growth through 2017/2018 did not lead to a large fall in unemployment as participation lifted from an October 2016 low of 64.4% to a January 2018 high of 65.8%. In the September Labour Force Survey, the fall in the unemployment rate to 5.0% was supported by a sharp 0.24ppt dip in participation to 65.44%. The all time record high in participation was 65.79% in November 2010. It recently hit 65.78% in January 2018.

There are two factors to consider in regards to participation. The structural longer run issue, such as rising female participation, education, aging population etc and the short-run dynamics where participation rises and falls with employment. In the long-run the cross currents of rising female participation and declining male participation (associated with an aging population - for now females are pushing out their age of retirement so the aging population is not a net negative for them) suggests we may have past a medium term peak in participation. Our forecasts focus on the short-run dynamics of participation associated with employment bounded by our view we have seen a structural peak.

A further reason to be cautious about lifting the participation rate back to record peaks is the observation that the recent improvement in male employment has not produced the lift in male participation that it has in the past. In fact while annual growth in male employment has lifted, the annual change in male participation has moderated. 
On this basis the participation rate drifts up a little through to the first quarter of 2019, before it drifts down a little as employment slows, limiting the rise in unemployment to 5.3%. Through the second half of 2019 and into 2020, employment lifts to an around trend pace and the participation rate continues to drift higher but remains short of the most recent peak of 65.78%.

Unemployment set to drift down to less than 5% by end 2019 while the rise in employment to population will remain shy of a new record high 

Once you have an employment and participation profile you can generate a profile for the unemployment rate and, as a cross check, an employment to population profile (including population estimates). Our forecasts result in the unemployment rate rising to a peak of 5.3% in the June quarter of 2019 before easing back to 5.0% year end then trending down to 4.8% by end 2020. The employment to population rate dips a little from 62.2% to 62.1% by the June quarter 2019 before entering a trend rise to 62.6% by end 2020. While not returning to the previous peak of 62.9% (August 2008) employment to population rises over this period highlighting that an employment pace of 2% is slightly faster than our current estimate of population growth.

Justin Smirk is a Senior Economist for Westpac