Here's our summary of key economic events overnight that affect New Zealand with news real wages are falling in Australia.
[Please note, there will be no Economy Watch podcast today due to some equipment failure.]
But first, after falling almost -20% over the past three weeks, US mortgage applications rose +2.8% last week on stronger refinance activity. Demand for mortgages to buy a new home actually fell last week.
We also got US housing starts data overnight, but for December. They rose +6.2% from November to an annualised rate of 1.4 mln, up from 1.322 mln in the previous month and well above the expected 1.33 mln. This latest data is the highest level since July and up from October’s 15-month low. Single-family housing starts rose +4.1% while multi-family starts climbed +10.1% to a three-month high.
New orders for US-made durable goods fell by -1.4% in December from November, following an upwardly revised +5.4% jump in November. The drop was milder than the -2% decline expected by analysts and was largely driven by a -5.3% fall in aircraft-related orders. Orders also declined sharply for capital goods, also largely aircraft-related. This leaves capital goods orders up +21% from a year ago, however, and overall orders up +12% on the same year-ago basis.
More currently, January US factory production rose +0.6% to be +2.4% higher than year-ago levels.
Interestingly, tariffs haven't been enough to stop the US losing another aluminium smelter, leaving the country with just five primary metal production plants.
And in other tariff news, Kevin Hassett, a top Trump economic adviser lashed out at a study from the New York Fed that found that American companies shoulder most of the costs of tariffs. He called for the central bank to punish the four researchers behind the work.
In Japan, manufacturer sentiment rose to extend the current positive run to 11 months. The Reuters Tankan index rose to +13 in February from +7 in January, driven by sharply better machinery orders - and a weaker yen.
In Australia, the Westpac–Melbourne Institute Leading Index six-month annualised growth rate, which indicates the likely pace of economic activity, stalled in January from December's positive outlook. This means they no longer have above-trend growth; it is back to their normal level.
And yesterday, the latest data out of Australia on wage growth pegs it at 3.4% in 2025. While that may seem a healthy rise, their CPI inflation rose 3.8% in the same year. Real wages are going backwards, it seems.
The UST 10yr yield is still just under 4.08%, up +3 bps from this time yesterday. The key 2-10 yield curve is unchanged at +62 bps. Their 1-5 curve is holding at just over +15 bps (-1 bp) and the 3 mth-10yr curve is just under +39 bps (up +3 bps). The China 10 year bond rate is unchanged at just on 1.81%. The Japanese 10 year bond yield is holding at 2.14%. The Australian 10 year bond yield starts today at 4.73%, up +4 bps today. The NZ Government 10 year bond rate starts today at 4.40%, down -5 bps from yesterday.
Wall Street is firmer with the S&P500 up +0.9% today on a rebounding tech sector. Overnight, European markets were up +1.2% except Paris which rose +0.8%. Yesterday Tokyo closed up +1.0%. Hong Kong, Shanghai and Singapore were all on holiday for CNY. The ASX200 ended its Wednesday trade up +0.5%. But the NZX50 ended up +1.7% and the best of the markets we follow.
The price of gold will start today back up +US$149 from yesterday at US$5007/oz. Silver is up +US$5.50 at US$78/oz today.
American oil prices are up +US$2.50 at just over US$64.50/bbl, while the international Brent price is now at US$69.50/bbl.
The Kiwi dollar is down -40 bps against the USD from yesterday, now just on 59.9 USc, a -5% devaluation. Against the Aussie we are down -60 bps at 84.7 AUc. We are down -10 bps against the yen. Against the euro we are down -20 bps at 50.9 euro cents. That all means our TWI-5 starts today down -40 bps from yesterday, now at 63.4.
The bitcoin price starts today at US$67,150 and down -0.6% from this time yesterday. Volatility over the past 24 hours has been modest at just under +/- 1.2%.
Daily exchange rates
Select chart tabs
The easiest place to stay up with event risk is by following our Economic Calendar here ».
19 Comments
If your job involves engaging logic, I wouldn't worry about AI taking your job tomorrow:
“The car wash is 40m from my home. I want to wash my car. Should I walk or drive there?” the prompt reads, with the distance varying among users, but still walkable.
AI often suggests to walk to the car wash. Better for you health, faster, saves wear and tear on the car.
https://cybernews.com/ai-news/ai-car-wash-test/
Most the newer AI models got it right.
This is always going to be the issue with AI, it can and does get things wrong, just like people. If a person gets something wrong the company can blame the person. If the AI gets it wrong, and the AI is provided by the company, then the company is at fault. Imagine if it gives incorrect legal advice for example.
Yep.
Planes pretty much fly themselves and have done for a while.
Yet we still demand human oversight.
And if there are no people involved in the next step to provide that logic?
It's easy trip up AI models if you use contrived prompts. The aim is to train them for everyday users who ask sensible questions.
Mmmm, have you ever dealt with many AI chatbots for customer service?
They are awesome at regurgitating answers you could otherwise gleam from the companies website.
They really struggle comprehending your question if it's outside of those rails. So you try to rephrase the question a few times, and then they say they'll need to refer you.
They've come along way in just 12 months. Another couple of years and they'll be reliably capable of most things.
For hundreds of billions of dollars, you'd like to think so.
There's definitely benefits, but also a great deal of futurism.
When the price of anything falls to zero demand tends to rise...
AI will allow us to build free software but someone will have to specify the problem and test cases that must be solved.
With AI we have never been able to build the wrong thing this fast, the key skill will be understanding the customer problem that needs to be solved and being able to articulate that.
It was crap last year, is stunning but still not quite there now, this is not moving in a linear fashion.
People who simply co-ordinate the movement of information across companies will be the first to go.
It was crap last year, is stunning but still not quite there now, this is not moving in a linear fashion.
What they are currently developing is mimicking human brain process observed in speech.
This is not all that's required to actually think.
So it'll process things, and possibly well, and fast.
But it's probably the wrong model to replace all human brain processes.
I would suggest that AI as it is being developed is like talented people. Each AI model or application will be good at something, perhaps more than one thing, but never everything. And like people will need to be reviewed, and that will take human involvement. Technology is never and cannot be fool proof. Everything breaks sooner or later.
The aircraft analogy further up the stream is point in context. Air France lost an Airbus into the middle of the Atlantic years ago. The final analysis determined that the air data probes had iced over (the primary fault) and neither the computers nor the pilots were able to figure out that the aircraft had stalled and they flew into the water from over 30,000 ft, fully stalled. Middle of the night, no visible horizon, but all the pilots needed to do was override the computer and push the nose down to get the aircraft flying again. Nightmare scenario and it would have taken some time for them to fall to the water.
It's like any trade or skill.
You can read the user manual, go to classes, watch a YouTube clip.
But there's "knacks" you can only learn via experience, as well as being able to understand the full context of the task you are undertaking.
Some can be paint by numbers, which you can get a machine or computer to handle.
Learning via experience is another way of saying that you've made all the mistakes and learned from them. Learning from shared knowledge might be less effective for learning but more cost-effective from a businesses pov
The 'free' software comes with caveats. How easy will it be to enhance and maintain going forward?
While AI assisted 'vibe coding' can produce working software very quickly, it's use beyond prototyping in a full production environment is yet to be proven
He points to the 76 percent of businesses surveyed for the company's latest Salary Guide who say they are planning to hire this year, up from 66 percent last year.
A quote from this article:
https://www.rnz.co.nz/news/thedetail/587245/reality-bites-for-job-seeke…
My question is, how many of that 66% from last year actually went on to hire more people?
It's this kind of stuff that is constantly printed that really grates on me. It's all about intentions with no trailing data to back it up. The recent unemployment release would suggest that few of that 66% ended up increasing their staff.
If you are trying to predict the future, then intentionality is one of the only ways you can do it.
Running a few businesses, I would say the trading environment has improved over 12-18 months ago (my sales confirm this). So, assuming things hold steady, or continue improving, I will need more staff (I have already hired a couple).
But I can't guarantee you what my figures will be in 12 months time.
So President Trump’s messenger wants the messengers of tariff reality, punished. Immediate vision of Bandinelli’s famous statue, Laocoon and Sons being strangled and devoured by snakes for daring to warn the Trojan hierarchy about dragging in that treacherous Greek horse.
Maybe worth noting:
https://timesofindia.indiatimes.com/india/indian-coastal-authorities-se…
India is now targeting the oil shadow fleet (a fleet of oil tankers operating amoung sanctioned states).
India has for the last 4 years been using the shadow fleet to obtain cheap Russian oil. Targeting the shadow fleet indicates India is now abandoning this strategy.
It also implies the shadow fleet are going to find it harder and harder to operate.
Russia lacks any credible naval capacity to secure shipping lanes for its oil.
So the resources are of little value if you have no way to sell them.
Recall enhancing pipeline supply? There was a mutual development for this with Pakistan for instance.Long distance and some challenging terrain though.
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.