Jump to-
Recent key developments in MAKE│NZ
•With it being such a busy time of year we’re afraid we’ll have to keep you in anticipation of interviews with our remaining two directors…They do say absence makes the heart grow fonder, so get ready to share that fondness with the time comes.
•Making Sense of a Confusing Political World*
In a year when headlines outpace history, ExportNZ, the NZ-US Council, and ASB are pleased to host US-based political strategist and professor Dan Schnur to help cut through the noise.
In this session, Schnur will share insights on how the current U.S. political cycle and broader global dynamics affect international trade, business confidence, and the operating environment for exporters. ASB Bank will also be on hand to give us an update on economic conditions both at home and abroad for New Zealand exporters. Expect a fast, frank briefing on how to interpret global turbulence and what it means for New Zealand exporters!
Dan Schnur brings decades of experience at the heart of American politics, serving on four U.S. presidential and three gubernatorial campaigns, as national communications director for Senator John McCain, and as chair of California’s ethics watchdog. Dan is currently a professor at USC and UC Berkeley.
This will be hosted at The George (50 Park Ter, Christchurch Central City) and if you’re interested in attending you can register HERE
•Hardware Meetup NZ are returning to Christchurch!*
If you’re interested in aerospace, hardware, or NZ’s innovation ecosystem, this is an event you don’t want to miss. You can join in person or register to watch the livestream. They’ll be having three speakers – Mark Rocket (Founder and CEO of Kea Aerospace), John Mann (An aerospace engineer and project manager with Tāwhaki National Aerospace Centre), and a third mystery speaker yet to be announced. If this is something you might be interested in you can find tickets and register HERE.
* Please note we have been sponsored by these events in the form of free ticket offerings, but that is not why we’re promoting these events, rather because as MAKE│NZ it’s our job to share news of events that may be of interest to our community
Recent key developments in New Zealand
•The Government has asked two of its Ministries, Foreign Affairs and Trade [MFAT] and MBIE, to jointly prepare a document called Long-term Insights Briefing [LTIB] on the subject of New Zealand’s productivity in a changing world. How can we accelerate the growth of high productivity activities in the New Zealand economy? The purpose of the document is to “… set out choices for government to help new high productivity sectors grow and identifies how international trade and connections enhance productivity more widely.”
After some fairly extensive public consultation, the two ministries have recently published a draft briefing document that they are again seeking feedback on. As our members represent a significant part of the ‘high productivity activities in the New Zealand economy’, MAKE│NZ has provided feedback in person, and in writing.
The document concedes that “productivity is a broad topic. We have chosen to focus on a particular aspect: How can we accelerate the growth of high productivity activities in the New Zealand economy?”
There is a recurrent theme here that regularly appears when ministers and government officials talk about economic development: A focus on “high productivity activities”. Productivity is defined as gross value added (GDP) per hour worked. The first point worth noting, then, is that by that measure some economic activities are inherently more ‘productive’ than others (Data from STATS NZ):
| Industry (Sector) | GDP in year to March 2025 ($m) | No. of employees in March 2025 quarter | GDP per employee ($ ,000) |
| Manufacturing | 21,836 | 222,248 | 98.3 |
| Construction | 16,930 | 177,949 | 95.1 |
| Financial and insurance services | 16,138 | 71,016 | 227.2 |
We don’t have exact data on the hours worked per year in each industry, but for the benefit of the above comparison, assuming they will be the same will be good enough.
Thus, contrary to an intuitive understanding of productivity, here it matters as much what we do as how we do it. That means the all-so-popular cross-country comparisons of productivity make little sense unless they are corrected for the share different industries have in the countries compared.
It also means that, in theory, the easiest way for a country to raise productivity would be to direct investment into the ‘most productive’ industries at the cost of the least productive sectors. That is exactly the point the late Sir Paul Callaghan made in his most famous slide from 2009:
There is, however, a slight glitch in this as it mixes Revenue per Employee and GDP per Employee (capita). The main message still stands, however.
So, which are the high productivity activities for which the draft LTIB recommends to government “adopting a more deliberate and strategic approach to enabling high productivity activities will be critical for lifting New Zealand’s overall productivity.”? The document rather carefully avoids being too specific on that question, but does mention value-added food and beverage, high and medium-high tech manufacturing, and ICT.
Overall, the draft LTIB document doesn’t contain a lot of specific recommendations for government policy or investment decisions. That is probably an appropriate reflection of the fact that productivity improvements happen first and foremost in businesses in the private sector. They are driven by their need to remain locally and globally competitive.
The document repeatedly emphasises the importance of skills, innovation and sector R&D, education and the science system, and regulations when it comes to improving productivity. These are all areas where government is the most important actor, and where the performance of recent governments has left the current and future governments with plenty of ‘opportunities for improvement’.
•Siemens-Gamesa is the world’s largest manufacturer of offshore wind turbines and has a significant market share in land-based installations outside of China. Recently, the company has been plagued by technical and financial difficulties and incurred significant losses. In Q3/2023 the situation got so bad that banks became reluctant to provide further loans to the company’s parent, Siemens Energy. The German government had to step in and provide the company with a five-year credit guarantee over €7.5 billion. The company has since entered a path of recovery and was able to exit the credit guarantee ahead of time in June 2025.
While the benefit of the above appears to be obvious – Siemens Energy employs about 86,000 people in Germany alone – there are questions in principle about the benefits of such arrangements. A 2015 OECD survey of member countries showed that 23 of them had a Credit Guarantee Scheme [CGS], New Zealand and Australia weren’t among them. So, what’s the reason the majority of OECD countries have adopted a CGS?
There are well-established methods to evaluate the benefits of credit guarantee schemes and the evidence across OECD countries demonstrates (here, here and here, for example) that these schemes deliver proven benefits including improved access to credit, substantially reduced borrowing costs, significant employment and sales growth, and enhanced business continuity during economic crises. These benefits are most pronounced for SMEs, young firms, and sectors with limited collateral-generating capacity. However, the success and long-term value of these schemes depend critically on proper scheme design, careful borrower selection, and avoiding excessive scale that could distort the market by artificially reducing competition.
New Zealand does not have a CGS as such. The Treasury’s Export Credit Office has been offering export credit guarantees since 2001, initially focused on capital goods projects and later expanded to support a broader range of exporters. Between April 2020 and June 2021, and in response to the COVID-19 pandemic, Treasury ran a Business Finance Guarantee Scheme [BFGS],operating on a risk-sharing basis, with the government guaranteeing 80% of loan default risk and participating banks covering the remaining 20%. The scheme offered a total of NZ$6.25 billion in loans to eligible businesses (NZ$5 billion from government backing and NZ$1.25 billion from banks). Individual loans were capped at NZ$500,000 with a maximum term of three years, targeting businesses with annual revenue between NZ$250,000 and NZ$80 million.
By international comparison, the BFGS was smaller than similar schemes in most other countries, but relatively had a much better uptake than in most other countries:
In December 2022, Treasury released a copy of the evaluation of the BFGS, completed by BERL in June of the same year. The review found that the roll-out of the scheme went well: “The general consensus from lenders was that the BFGS achieved the objective of facilitating the provision of credit to cushion the impact of COVID-19 on otherwise viable SMEs facing temporary financial distress.” Not surprisingly, there was room for improvement, and the review did not provide any estimates of the macroeconomic impact of the scheme.
As we mentioned earlier, the Labour Party recently announced plans to set up a Future Fund, the aim of which is to “invest in infrastructure and innovative Kiwi business”. Based on the evidence from other OECD countries, it may be a better idea to set up an enduring Credit Guarantee Scheme alongside the existing Export Credit Guarantees, rather than using taxpayer funds to take an equity stake in individual businesses, with the former providing a better return on investment to the economy than the latter.
Recent key developments in the World
•Currently the focus of AI adoption in business is on routine cognitive tasks associated with the processing of text, images/video and sound, slowly expanding into non-routine cognitive tasks as well
•Consequently, the use of AI is limited in volume and in the number of tasks where it can gainfully be deployed in manufacturing
•There are future opportunities, however, with the development of AI agents tasked with automating specific process control and process management functions.
•In a recent conversation with a manufacturer about the application of AI in manufacturing, his comment was, somewhat dismissively “it’s all about robots”. In reality, most – and the most productive – application of AI doesn’t involve the replacement of human activity in physical processes by machines (process automation). Rather, it involves the automatic capture and collation of (shop-floor) process data, and the extraction of information from that data. Predictive maintenance is the most commonly mentioned example, but also quality control using digital image analysis, and process optimisation. The next iteration will be the use of AI agents. An SME may use a platform that integrates a Large Language Model (LLM) with its Manufacturing Execution System (MES), enabling operators to ask natural language questions like, “Why did production line 3 have a lower output yesterday?” The AI will then analyse maintenance logs, sensor data, and quality reports to provide a summarised root cause. MES provider Tulip , for example, is working on offers along these lines already.
The drive towards the use of AI at least in larger SMEs is driven by:
- Cost of Sensors and IoT: The price of sensors, gateways, and data connectivity has plummeted, making data collection affordable.
- Cloud Computing: Eliminates the need for massive upfront investment in on-premise servers and IT staff.
- Proven ROI: There are case studies available now that create a compelling business case. Reducing downtime by even a few percent can save an SME hundreds of thousands of dollars.
- Competitive Pressure: As larger suppliers adopt AI and become more efficient, SMEs in their supply chain are pressured to follow suit to remain competitive.
However, there still are significant barriers:
- Lack of In-House Expertise: The primary barrier for SMEs is the skills gap. They often lack data scientists or AI specialists.
- Data Readiness: Many SMEs have “data silos” or legacy machines that are not connected. The first step is often a digitalization project, not an AI project.
- Cost and Perceived Risk: While costs are falling, the investment is still significant for a small business, and the perceived risk of a failed project is high.
- Cultural Resistance: A shift from traditional, experience-based decision-making to data-driven, AI-informed processes can face internal resistance.
•A few weeks ago we talked about youth unemployment globally, and the role of AI. The number of articles reporting on the impact of AI on work keeps growing. However, these are usually based on theoretical studies or on intention surveys, rather than actual data on AI at work.
Fortunately, there is now also a small number of very thorough empirical studies on the topic, reporting on what is actually happening in labour markets that can be associated with the introduction and use of AI – see here, here and here.
These papers are based on large sets of data and contain information on many aspects of labour market dynamics.
We’ll limit ourselves to aspects related to manufacturing. One paper reports on the Top and Bottom 40 occupations where exposure to either ChatGPT or Claude LLMs is significant and finds that among the latter the vast majority of occupational activities involve manual labour and/or a physical interaction with machinery and equipment. The Top 40 group is dominated by occupations involving the processing of verbal (text), visual (images) or acoustic (music) information, or a combination thereof.
For their analysis, the authors ordered occupations into four main groups: Cognitive (routine, or non-routine), and Manual-Physical (again routine, or non-routine). They find the biggest impact of AI in non-routine cognitive tasks, with some impact also on routine cognitive tasks.
Routine, and even some non-routine manual-physical tasks have been strongly affected by automation and robotics prior to the arrival of AI, but the authors conclude that “despite AI’s rapid advances in language, reasoning, and pattern recognition, current systems lack physical embodiment and thus aren’t suited for tasks requiring spatial awareness or physical exertion. These results suggest that occupations involving manual labor remained relatively insulated from AI-driven disruption as AI capabilities advanced from pre-LLM models to multimodal models. However, this insulation may diminish over time, as AI capabilities are increasingly integrated into robotics platforms.” A review of the state of the art for the latter can be found here.
A recent paper by the Reserve Bank of Australia showed trends in the Australian labour market based on this classification:
Another of the papers quoted above takes a more granular approach to study the impact of AI on a range of occupations, using heat maps to illustrate their findings:
Again, for sectors (industries) we find that the use of AI in manufacturing is comparatively limited and mainly focused on Modelling (digital twins?), Robotics, Image Analysis and Process Optimisation. For occupations, and focusing on Production, AI is used most in Task and Workflow Automation, and Process Optimisation:
Finally, these authors also looked at the relationship between company size and AI utilisation and, for manufacturing, confirmed relatively low levels of utilisation and the same concentration in larger companies observed for other sectors:
•And, finally, and still on AI: How do you recognise a truly great intellect at work? When (s)he repeatedly says “I don’t know” in response to a difficult question. Watch this interview with Geoffrey Hinton, who was awarded the Nobel Prize in Physics in 2024 “for foundational discoveries and inventions that enable machine learning with artificial neural networks”.
Leave a Reply
You must be logged in to post a comment.