Three Non-Consensus Ideas for 2025

As the world changes, it is tempting to try and build an investment process or a business on shifting sands. We always remind ourselves of the Jeff Bezos quote from 8 years ago:

“I very frequently get the question: ‘What’s going to change in the next 10 years?’ And that is a very interesting question; it’s a very common one. I almost never get the question: ‘What’s not going to change in the next 10 years?’ And I submit to you that that second question is the more important of the two — because you can build a business strategy around the things that are stable in time. … In our retail business, we know that customers want low prices, and I know that’s going to be true 10 years from now. They want fast delivery; they want vast selection. It’s impossible to imagine a future 10 years from now where a customer comes up and says, ‘Jeff I love Amazon; I just wish the prices were a little higher,’ or ‘I love Amazon; I just wish you’d deliver a little more slowly.’ Impossible.”
– Jeff Bezos, Amazon founder, July 2016

Benchmark hugging large cap portfolios were a thing in the early 2000’s – doing what some small cap funds are doing today in trying to manage business risk and client risk by becoming overly concerned about relative to index risk. Then came the misguided era of active share – where highly concentrated portfolios not only overcame the complaints of benchmark hugging, but also had some timely (but ultimately survivor biased and as such flawed) academic research that pointed to this being a superior approach. And then finally we have technology itself, the ultimate ever changing silver bullet which offers a series of truly transformational opportunities and many more red herrings that can tempt investors and managers into new paradigms.

Like Mr. Bezos, we try and stay focused on what is true and stable over time for our clients and work backwards from there. For Longwave, that is a view that clients today and in the future are likely to value More Consistent Alpha at Lower Fees. As it relates to issues of benchmark awareness, we will happily hold no position in a large benchmark weighted stock. As it relates to active share and concentration, we are happy to diversify our portfolio into many high-quality businesses rather than limit to a small arbitrary number. And as it relates to technology, we have always been progressive in using any technology that helps us achieve our goals, without going to either extreme of having to adopt every fad, or conversely becoming technology luddites and forsaking real productivity and investment insight benefits.

Below are three specific ideas for 2025 where Longwave may take a slightly different-to-consensus view.

It seems to us the active small cap market has become far more benchmark aware compared to 20 years ago, holding names close to benchmark weight to reduce the pain of relative performance if the stock goes up a lot. This never used to be a thing. When building a portfolio of 30-40 stocks from an opportunity set of 1,000, a slot was never used to hedge index risk. Indeed having this many stocks to choose from and building a portfolio of 100-120 names still allows us to only invest in the stocks that meet our investment criteria.

But we get it. We feel the pain. Most of our underperformance relative to the benchmark and peers in the past two difficult periods (2020/21 and 2023/24) came from names we didn’t hold, that were significant weight in our benchmark, and that went up a lot. This cycle (2023/24) think Life360, Zip Co, Ventia, Judo Capital, Mesoblast, Paladin Energy. Paladin had the wonderful pattern of doing all its speculative work upwards in the small cap index, then being promoted into the ASX100 before reality – and a 60% share price fall – set in. It is this pain of relative underperformance a benchmarking strategy is trying to protect against.

Last cycle (2020/21) Pilbara Minerals, Lynus Rare Earths, Allkem, Galaxy Resources, Megaport, Lifestyle Communities were the culprits. But we would also observe the duration of this pain (1-2 years) is much shorter than the investment lifecycle of our strategy and the time horizon of our clients (5-7 years or more). For our investment approach in particular, this short-term pain really only shows up in relative performance terms, as absolute performance over both these periods was very high due to euphoria enabling many low quality stocks we shun to soar. Given the long-term performance we have achieved, it is a short-term price we are willing to pay.

Over the past 15 years, the need for active share to be a fund feature clients should desire was founded on faulty research. This has sent portfolio construction innovation in the wrong direction and has probably been net negative for public market alpha.

Higher active share leads to higher alpha – both positive AND negative. The highest returning funds are most likely to be the highest active share funds. BUT the lowest returning funds are also most likely to be the highest active share funds. Nothing too controversial here.

The big problem was when the original active share study (How Active is Your Fund Manager? A New Measure That Predicts Performance – Cremers and Petajisto, 2006) was done, it used survivor biased data. Not surprisingly, if you are a fund manager who has underperformed by 4%, 7% or 10% per annum, you eventually cease to have any money to manage and drop out of the surveys.

Outperformers of this magnitude suffer no such fate and survive – hence the bias.

If you remove the dead managers, you can find a positive relationship between active share and positive alpha. But that is not how the world works – when we invest today, we do not know which side of the performance distribution our high active share manager will be.

Active Share research has failed to replicate and been disproven multiple times since publication; however, it is such a compelling idea for active managers – who have little good news to point to these days – its mythology is still repeated almost 20 years later.

We have previously expressed skepticism at the way the market extrapolates and values specific trends in technology – which are exponential and highly uncertain. Our prior concern was about the pricing of a hot market thematic (data centres = AI) not the utility of the new technology itself.

As it applies to our own work, we have been experimenting with AI models and tools of various sorts during 2024 and are likely to put some to work in 2025. We are improving our understanding of technologies, how they are changing, and discovering real use cases this technology can provide – which was not available with prior technologies – on both sides of our investment process.

The standard line when talking about AI is it won’t replace the need for human decision-making in the investment process. This may be true, but it may change the number of humans needed, how they spend their time, or the make-up of skills across an investment team. Or maybe it won’t. But AI technology itself is so transformative and advancing so quickly on both performance (improving) and cost (plummeting) that we believe priors and biases need to be continuously challenged.

In the past 20 years we have witnessed and embraced a technological evolution for the generation of insight from fundamental active management. We believe AI will likely again change fundamental active investing, but this is not necessarily a shared belief across the market.

At Longwave we have taken advantage of several new technologies in the past two decades (here and at predecessor firms) which have improved client performance and made our investment process more consistent and more efficient. A considered implementation of technology has allowed Longwave to build an investment philosophy around a core client value: deliver more consistent alpha at lower fees.

Our experience does not reflect a universal observation across our industry. Many investment teams run an investment process that when viewed through the lens of technology could be considered state of the art circa 2004. There have also been fads (like Big Data) that have not changed core research processes, and for many investments teams AI is perhaps seen as nothing more than the latest and inconsequential fad. Our view entering 2025 is that AI will be a highly consequential and enabling technology for fundamental investment research.

What is our goal of using AI in our investment process? Well for fundamental research outcomes, we believe it could significantly increase the quantity of research at the same or better quality per investment professional. This means fundamental analysts can cover more stocks with no reduction in the speed and quality of insights, or investment teams with smaller universes (like large cap Australian equities) can cover their opportunity set with far fewer people – with obvious cost and fee implications, hopefully some of which is passed onto their clients.

For many years, the arms race of headcount in investment teams (more people = better) was the conventional wisdom. In most other industries, the use of technology to simultaneously improve outcomes and reduce costs is not a controversial goal. AI will likely bring that thinking to some corners of the fundamental investment management industry.

“The future is already here – it’s just not evenly distributed.”
― William Gibson

On the investment front, innovation from AI may produce winners – but finding those early is fraught. Disruption from AI will produce losers – after all as Jeff Bezos also said, “your margin is my opportunity” and pointing AI technology at existing profit pools is a good place to seek early returns. We are looking for the next Fairfax Media – a company generating huge profits and trading at a 20% premium to the market at the time the Internet came about. 20 years on and surviving media firms (Nine Entertainment Co and Seven West Media) have had profits go backwards (significantly relative to market earnings growth) and relative to market P/E multiples are now at 30-60% discounts.