Days to the General Election: 28
See Party Policies here. Party Lists here.

Rodney Dickens says the risk isn't that economic & employment growth will deteriorate, but rather low rates & high migration will drive growth in spending, GDP & jobs

Rodney Dickens says the risk isn't that economic & employment growth will deteriorate, but rather low rates & high migration will drive growth in spending, GDP & jobs

By Rodney Dickens*

At times the economic forecasters say things that are dubious at best, which sets off the bullshit meter in my head.

Normally I have a chuckle to myself and get on with life providing clients with quality insights into economic and property market prospects.

But the comment by a Westpac economist that "Strong net migration in the near term is one reason we expect unemployment to rise over the next year" made me chuckle enough I thought the joke was worth sharing.

Suggestions that strong net migration (i.e. strong net immigration) will contribute to the unemployment rate increasing from 5.9% currently to 6.5% over the next year will be music to the ears of the anti-immigration lobby, but the opposite is likely to happen.

What is the normal link between net migration and the unemployment rate?

Is it normal for high net migration (i.e. immigration less emigration) to be associated with a rising unemployment rate?

Prior to the current one there have been two previous major upturns in net migration since 1990 and in both instances they were associated with a falling unemployment rate and the periods of peak net migration were associated with or followed by low unemployment rates (adjacent chart). Setting aside the increase in the unemployment rate in the last two quarters that may reflect sampling errors, it has fallen during the latest upturn in net migration. Off goes the bullshit meter in my head.

The Westpac economist went on to write: "over the next year we expect the unemployment rate to rise to 6.5%, as slower jobs growth meets continued rapid population growth" driven by high net migration.

If I assumed weak economic and employment growth in the face of robust growth in the labour force driven by high net migration I would also be predicting a rising unemployment rate. My real beef is with the Westpac economists predicting slower GDP and employment growth in the face of the stimulus from strong population growth, the sizeable fall in interest rates over the last 15 months and a significant fall in the exchange rate. The left chart below shows the predictions for annual GDP growth by the Westpac economists, while the right chart shows their latest predictions for annual employment growth.

The Westpac economists are predicting a larger slowdown in annual employment growth than predicted by the tumble in the ANZ survey of employment intentions (adjacent chart). This is puzzling for a number of reasons.

The ANZ business survey has gone AWOL as an economic indicator, as covered in a recent Raving. This applies to the employment intentions survey and most of the components of the ANZ survey and reflects some respondents ticking the negative box to provide Governor Wheeler with the evidence of weaker economic growth he is seeking to justify OCR cuts. It would appear that the economists at Westpac have been sucked in by this game by more than 100% (i.e. they are predicting that employment growth will fall by even more than the dubious fall in the ANZ survey of employment intentions).

What evidence do I have that the fall in the ANZ survey of employment intentions largely reflects a game being played by some respondents more than a deterioration in employment growth prospects? Firstly, while the ANZ survey of employment intentions has tumbled, noting the ANZ survey is long standing and gets lots of media attention, the less well publicized and seemingly uncorrupted BNZ Business NZ survey of employment intentions remains not far below peak levels (left chart below). Secondly, while the ANZ survey of business confidence has tumbled into negative territory, predicting that annual GDP growth will soon turn negative, as the survey did incorrectly during much of the time Labour was in power (right chart below), the BNZ Business NZ surveys of the service and manufacturing sectors, when combined, point to a near-term mild improvement in annual GDP growth (2nd left chart below). Thirdly, our Growth Driver Index is predicting stronger not weaker GDP growth (3rd right chart below).

The SRA Growth Driver Index quantifies what the key drivers of economic cycles are predicting (i.e. interest rates, net migration, the exchange rate, export prices and petrol prices) and has a dramatically better track record than the economic forecasters, especially at predicting stronger economic growth, as it is at the moment. As outlined in our pay-to-view economic reports, the SRA Growth Driver Index gets ahead of itself in predicting major slowdowns in economic growth, as highlighted by the two boxes in the right chart above. This is because when low interest rates combine with high net migration they fuel speculative bubbles in the housing market, like the unfolding one, which makes the housing market and GDP growth resilient for a period in the face of subsequent interest rate increases that drive the SRA Index down. The risk isn't so much that economic and employment growth are about to deteriorate significantly as the Westpac economists are predicting, but rather that, as occurred in the mid-1990s and again in the mid- 2000s, the combination of low interest rates and high net migration will drive robust growth in consumer spending, GDP and employment (see this link for the strength in electronic card spending reported by Statistics NZ for September). 

This has major implications for a wide range of areas, including the housing market, residential building, consumer spending growth, employment growth, interest rates and the exchange rate as covered in our various pay-to-view reports. Now more than ever people making important business and/or investments decisions need quality insights, like those provided by our pay-to-view reports, rather than some of the bullshit-meter-ringing garbage dished up too often by the economic forecasters.

Always check both scales on a chart

The report by the Westpac economist on migration prospects that included the reference to a rising unemployment rate had some useful content (i.e. the bank economists don't always or entirely dish up bullshit). The economist commented that: "Statistics New Zealand doesn’t break down migrant departures by visa type, but historically about a third of foreign migrants (excluding the small number of Australians, who don’t need a visa) have eventually left, with the number of departures spiking after about two years." - that is, the surge in foreigner immigrant arrivals over the last two years should contribute to more departures by foreigners resident in NZ over the next two years, contributing to lower net migration.

The left chart below is a replica of the chart in the report demonstrating this point, with it showing immigrant arrivals of foreigners (right scale) as a useful leading indicator for the number of foreigners resident in NZ leaving on a permanent or long-term basis (left scale). The highest correlation, at a respectable 0.88 (like an 88% mark in an exam), is with arrivals leading departures by 28 months, in line with the "about two years" identified by the Westpac economist. An inspection of the chart without checking the scales gives the impression there is about to be a dramatic outflow of foreigners who arrived in the country over the last two years. In defence of the Westpac economist, he did make reference to a three for one relationship, but this is easily missed if you, like me, are inclined to skip the text and focus on the charts.

The right chart below replicates the left chart but uses the same scale for permanent and long-term arrivals and departures by non-NZers. It shows that arrivals, that have increased further recently, are dramatically more cyclical than departures (i.e. based on past experience the likely, imminent upturn in departures by foreigners resident in NZ should be relatively modest in the context of the major cycles in net migration). The anti-immigration lobby can probably look forward to two disappointments. Their cause probably won’t be aided by high net migration driving up the unemployment rate and there is likely to be only a relatively small increase in the numbers of recent immigrants leaving compared to the high numbers arriving, although there is a case for expecting some fall in the numbers foreigner immigrant arrivals, as discussed in our Housing Prospects and Building Barometer reports.


*Rodney Dickens is the managing director and chief research officer of Strategic Risk Analysis Limited.

We welcome your comments below. If you are not already registered, please register to comment.

Remember we welcome robust, respectful and insightful debate. We don't welcome abusive or defamatory comments and will de-register those repeatedly making such comments. Our current comment policy is here.


Rodney clearly is very pleased with himself and his charts. But his charts looking back don't show any previous circumstance like the current one in Auckland. We have had price rises and falls before, but nothing like this current speculation bubble.
Migration and other causes of demand drive the rental market, which are curiously this time quite disconnected from selling prices. Indicating a bubble. Rodneys charts are inadequate for that.

Great charts wish they went back further.. also need some charts showing ecological damage, social cohesion and quality of life.. because as the population density rises NZ becomes more violent, congested and unpleasant to live in. Perhaps a graph showing infinite growth against a background of dwindling resources on a small pacific Island. But as long as it keeps wages low and rentals full its all good and the future looks bright.

This raving has my bullshit meter ringing at full bore!!

What Rodney has absolutely failed to recognise is that both of the past two immigration booms were DURING periods of declining unemployment ie periods with plentiful work. The latest immigration boom is during a period of reasonably stable employment. That is the difference.

The bank economists may not be 100% right but they are definitely on the right track.

Rodney's bullshit meter needs a reboot...

What about the risk to our declining GDP per capita?

The analyses are interesting. It would be useful to identify the time-lag in the Unemployment/Net Migration Chart and try to link this with external factors such as the conditions in the origin countries of migrants. Likewise, how much of the time-lag between arrival and departure is affected by purpose of migrating: was it because of short-term intentions; of failed expectations; or cultural reasons? That sort of information is probably not readily available..

I am not sure what the "Correlation=0.88" means in the last graphs. Does it refer to the cross-correlation between arrivals and departures, in an endeavour to identify the time-lag, or does it refer to the 3-month smoothing, or to the seasonal adjustment? As matter of interest, how applicable is normal statistics to these numbers?

I appreciate that endeavouring to model information of this nature is extremely difficult but I think it is worth trying to, at the least, test the prognostications of others.

Rodney Dickens - I don't believe it

The first error is the assumption the unemployment data is true

Unemployment data is based on a selected sample of 15,000 people who are surveyed every month for 6 months, then the selected sample is replaced with a new 15,000 people. The sample will hopefully, randomly comprise a broad spread of occupations some of which might be exposed to new migrant pressures some won't. The chances of one particular sample being immune to migration changes are high. The chances of one 6 month sample being exposed to migration pressures are equally high, although unlikely. The time of greatest volatility occurs when the sample is changed

The second error is to graphically compare one set of figures which are actual, to another set of figures which are not actual but simply an unreliable sample survey

In 2014 the Australian Bureau of Statistics was ridiculed when, during a time of huge lay-offs, the number of new jobs created climbed by an unheard of 66,000. then the following month all the gain was lost, yet there was no announced sample change .. as a result, the ABS has had to change it's methodology. The ABS never announces when it changes the sample

The logic of that was the huge number of lay-offs did not appear within the sample survey and potentially would not appear in the next sample of 20,000 (AU used to use 40,000 which was cut to 20,000 as a cost-cutting measure)

Did not hear of any similar revisions in New Zealand

The day the ABS acknowledged its survey method of measuring employment was flawed ... ABS data disaster is final proof Australia needs new measure of employment ... Oct 10, 2014 - Yesterday Australians were told the ABS measures of employment and unemployment cannot and should not be 'trusted'. The current methods used by the ABS to measure the labour force and estimate unemployment and under-employment are based in post-World War 2 thinking – nearly 70 years ago. The questions used are flawed for many reasons – typified by saying a person who works just one hour a week is employed

google "abs error measuring unemployment"

This is the problem with trying to predict the future... Everyone has a different theory. And chances are everyone is wrong.

Exactly. Everyone is wrong most of the time. But knowing that you are wrong most of the time is probably going to give you the best chance of making the right decision.

Pardon? If I knew I was going to be wrong most of the time, why would I make a decision?

You make a decision based on it hopefully giving you the least wrong outcome!

Let's face it, if you decide to make an investment, it's highly improbable that it is going to be the best investment you could have made - some stock or property or whatever somewhere else on the face of the earth will almost certainly prove to be a better investment over the same timeframe...

So give up trying to get everything perfect and just get it the least wrong you can!

Making a not so good decision is better than making no decision. (It's a bit like an exam, you are going to answer all the questions even if you think some might be wrong (except in those school Maths competitions where you get penalised a quarter of the marks for choosing the wrong answer...).

Days to the General Election: 28
See Party Policies here. Party Lists here.