July unemployment rate likely ~5%

Posted by

My judgement is that the Australian unemployment rate will be steady, or up a touch to 5% in July.

Given that the RBA held rates at their August meeting, it seems fair to assume that the unemployment rate holds the key to the policy outlook — at least until the next +0.9%q/q core CPI result.

The RBA would probably like to see the unemployment rate drift up over 5% – as they have always had a problem with inflation when unemployment is sub-5% – however they won’t want to see it spike in any material way. If it did spike, the RBA’s inflation forecast would decline and they would be in a position to ease policy.

It was with some surprise that I read the Roy Morgan labour market poll results – which showed the unemployment rate +60bps to 7.6%.   The Roy Morgan measure is not seasonally adjusted, and there is a typically a hefty seasonal in July (say ~30bps) – however if this is accurate, it is not a good sign.

I doubt it’s correct. There was a slight increase in (my seasonally adjusted) Google searches for “unemployment benefits” – however it was modest, and is consistent with the seasonally adjusted unemployment rate remaining around the current 4.9%.

Of interest to me is that NSW has joined Qld as the most ‘searching’ State. Perhaps those NSW public sector redundancies got things moving …


  1. The participation rate has been falling in the last few months and that has kept unemployment at around 4,9% despite not many new jobs created (according to a survey with a *significant* margin of error). Is that a leading indicator at all or has no bearing on the main unemployment number? thanks

    1. There has been some funny business going on with the survey, it seems. So it is hard to tell. The hazard rates (using matched sample gross flows data to see when employed units lose their jobs) look okay so far, but that could turn.

      Empirically, the PR is pro-cyclical, and coincident – but you would want to see more data before saying that the recent move means anything, especially as the floods will have probably played some part in the sampling issues.

      Remember this survey is designed to measure one thing – the unemployment rate. The ABS thinks that the trend measure of the UR is pretty good.

      All other bits, especially month to month SA jobs, are weaker estimates. Particularly jobs given the tousle the ABS has had with population estimates. We should have an establishment survey for that, like the USA.

      Sent from my iPad

    1. They call X people, if Y are employable (= working or looking for work, employed or unemployed) then PR = Y/X

      1. Stratification etc aside, yes that is it. That is why the ratios are better measured than the levels numbers. The recent problem, as I understand it, is that a bunch of highly employed units moved out of the survey two months ago in northern NSW. This may have been flood related.

        Remember that the survey unit is a physical address – hence the title ‘household survey’.

        A related issue, and one the RBA touched upon, is that weak jobs growth is due to slow inflation of the ratios due to low population growth.

        Sent from my iPad

        1. Yes, the ratios are more reliable being the absolute numbers projected based on population estimates. They ain’t calling that many people either… so even the month to month ratios should not be relied upon too much…. Trend ratios are the significant ones.

          Regarding household survey, they seem to be calling a few businesses too:

          “The Labour Force Survey is based on a multi-stage area sample of private dwellings (currently approximately 29,000 houses, flats, etc.) and a list sample of non-private dwellings (hotels, motels, etc.), and covers approximately 0.33% of the civilian population of Australia aged 15 years and over.”

          1. Yeah, they survey hospitals and motels etc. These tend to have higher unemployment rates, which is why the gross flows ratios look better than the total ratios.

            Sent from my iPad

  2. Roy Morgan survey sees unemployment ticking up:

    “The July 2011Roy Morgan employment estimates was down 418,000 in a month to 10,802,000 with the falls consistent across both full-time employed — 7,421,000 (down 219,000) and part-time employed — 3,381,000 (199,000).

    “This month’s fall in employment followed four straight months of rises and takes Australian employment back to the levels of February (10,770,000) and March (10,800,000) as school-leavers began to join the workforce after Summer holidays at the beginning of the new financial year.

    “This is the biggest monthly fall in Australian employment ever recorded by the Roy Morgan employment measure and shows that the RBA’s decision to ignore commentators calling for an interest rate rise this week was the correct one. Unemployment now stands at 7.6% (885,000, up 40,000) while underemployment is virtually unchanged at 7.3% (859,000, down 3,000). This is an overall level of unemployment and underemployment of 14.9% (1,743,000, up 36,000).

    “In addition employment in July is now down 106,000 on a year ago in July 2010. This is the first time employment has fallen year-over-year since October 2009 (10,385,000) which fell compared to October 2008 (10,400,000) — when the Global Financial Crisis began and dropped employment.”

    If employment cracks, and unemployment starts rising rapidly (no amount of mining boom employment will offset the crash in retail) then the last pillar of the Bullhawks’ arguments will be undermined, and Bill Evans will be declared an economic Nostradamus.

  3. Great stuff. I haven’t seen updated Taylor rule from you for a while. Do you have one handy?

Comments are closed.