iShares Robotics and AI ETF IRBO vs. ROBO Global
When it comes to investing in the many technology themes we cover here on Nanalyze, we’re always quick to recommend that investors diversify instead of trying to find the next Microsoft. Oftentimes, that diversification comes in the form of Exchange Traded Funds (ETFs). The simple way to explain an ETF is that it’s like a mutual fund (basket of stocks) except with lower fees. ETFs also typically track an index, so you don’t have to worry about some overpaid fund manager underperforming the market and then taking your money anyway.
When it comes to ETFs, iShares is the king of ETF providers with more than 800 ETFs globally accounting for over $1 trillion in assets under management (AUM).
iShares is a subsidiary of Blackrock (BLK), the world’s largest asset manager with $6.3 trillion in AUM, and they have now decided to launch a fund that provides exposure to robotics and artificial intelligence. It’s called the iShares Robotics and Artificial Intelligence ETF (IRBO), and it launched in July of this year having gathered a meager $10 million in assets so far. The impetus for this launch was likely the success that ROBO Global has had with their ROBO Global Robotics and Automation Index ETF (ROBO).
We’ve written about The ROBO Global Robotics and Automation Index ETF before, and the product is a clear success today with around $2 billion in assets under management. The company that built ROBO has been growing even bigger with new ETF launches covering different segments and variants of their investment universe. iShares’ decision to build their vehicle instead of buying into an existing methodology is a clear challenge. Let’s take a look at how the two offerings differ.
ETFs, with very few exceptions, are based on indices representing a certain regional, thematic, or style segment of the stock market. These funds then aim to track their chosen index as closely as possible, with as low a cost as possible. Hence the choice of underlying index, and its calculation methodology, will define the exact exposure investors will receive from the fund. ROBO uses its own Robotics and Automation Index which they put together using a team of PhDs who select stocks to be included in their index from “a proprietary database of Robotics and Automation Companies“. IRBO uses the NYSE Factset Global Robotics and Artificial Intelligence Index which is rules-based. These two indices have significant differences in terms of coverage, constituent selection, weighting, and rebalancing.
The ROBO underlier focuses on companies active in 12 sub-sectors across the robotics and automation value chain. Their methodology declares that eligible companies have to have at least a distinct portion of business revenue derived from robotics & automation, and also have potential to grow within their space– these are their so-called non-bellwether companies. Bellwether constituents are the ones we usually refer to as ‘pure-play’ and these derive most of their revenue from robotics and automation activities. Their method of determining which stocks fall under these categories is proprietary. In other words, you do not get to see how the sausage is made.
IRBO, on the other hand, recognizes 22 sub-industries across the robotics, automation, and AI value chain and requires constituents to derive 50%+ of revenues from these activities or be market leaders in any of the sub-industries. They use an industry classification standard called RBICS – a Factset industry classification system with up to three times the depth found in commonly-used industry classifications – and then apply the following rules:
You can now see exactly how this rules-based methodology works for IRBO. You will also see later just how much this rules-based methodology disagrees with a team of PhDs who are using a methodology with proprietary elements.
Size and Liquidity Criteria
Both indices define minimum size and liquidity requirements for company inclusion in order to ensure all constituents remain investable on the market. ROBO defines their minimum market cap as $200 million for newcomers and $100 million for existing constituents. IRBO’s size rule is stricter requiring $500 million and $400 million respectively.
Liquidity is measured by 3-months average value traded. ROBO requires new components to have an average of $1 million and existing ones $700,000. IRBO’s requirements are $2 million and $1.5 million for the same brackets. In addition to the above rules, ROBO allows 65 components as a minimum and 200 as a maximum while IRBO doesn’t put constraints on the number of companies included.
Weighting and Rebalancing
ROBO uses a modified equal weighted approach to determine component weights. Bellwether (or pure-play) stocks are set to 40% of the total at each review while non-bellwether stocks make up 60%. Components are equal weighted within these brackets. Index reviews are done quarterly, meaning the index follows actual markets closely but can incur more trading costs due to more frequent readjustments to the market. IRBO uses a simple equal weighting procedure done at each of the semi-annual reviews, resulting in relatively simpler rules and lower costs.
Top-10 Names and Overlap
Significant differences in the two funds’ index methodologies result in different member lists. Here is the comparison of their top-10 holdings taken from their latest published lists as of the 11th of September.
|Ticker||Company name||Weight (%)||Ticker||Company name||Weight (%)|
|AMD||ADVANCED MICRO DEVICES||2.19||AVAV||AEROVIRONMENT||2.81|
|DDD||3D SYSTEMS||1.47||BRKS||BROOKS AUTOMATION||2.01|
|PRLB||PROTO LABS||1.42||OII||OCEANEERING INTL||1.97|
|LSCC||LATTICE SEMICONDUCTOR||1.41||ZBRA||ZEBRA TECHNOLOGIES||1.93|
|2379||REALTEK SEMICONDUCTOR||1.4||ROK||ROCKWELL AUTOMATION||1.87|
Incredibly, there is only one overlap in the list of top-10 names, iRobot, the robotic vacuum cleaner manufacturer. You would think these ETF providers could at least agree on half the names in an “AI and robotics” top-10 list, but it’s actually much worse than that when we dig into the rest of the names. These two ETFs can only agree on 15 stocks between the two of them as seen below – along with links to the articles in which we’ve covered them:
|IRBO and ROBO Overlap|
|Ticker||Company name||Description||Weight in ROBO (%)||Weight in IRBO (%)|
|2049||HIWIN TECHNOLOGIES||Motion control products||1.15||0.71|
|6324||HARMONIC DRIVE SYSTEMS||Precision components for robots||1.26||0.89|
|6506||YASKAWA ELECTRIC||Precision components and fully built robots||1.30||0.8|
|6954||FANUC||Automation products and robots||1.65||1.14|
|CGNX||COGNEX CORP||Machine vision systems, software, and sensors||2.07||1.26|
|DDD||3D SYSTEMS||3D printers||1.37||1.47|
|IRBT||IROBOT CORP||Robot vacuum cleaners||2.80||1.72|
|ISRG||INTUITIVE SURGICAL||Surgical robots||2.03||1.25|
|MCHP||MICROCHIP TECHNOLOGY||Integrated circuits||0.84||0.91|
|NVDA||NVIDIA||Graphics processing chips||1.08||1.16|
|QCOM||QUALCOMM||Semiconductor and telecommunications equipment||1.24||1.33|
|SSYS||STRATASYS||3D printers and 3D production systems||1.20||1.30|
Perhaps a better way to view this is by looking at the overlap using a Venn diagram. Here we can see just how badly the two index providers disagree on just about everything.
For investors, this is quite disconcerting. If you had to pick one, which would you pick? Would you go with the world’s biggest ETF provider and trust their experience or would you follow the herd and invest in ROBO which has accumulated more than $2 billion in assets? Maybe buy both? The bigger problem is that these two firms can’t agree upon which stocks give investors exposure to AI and robotics – which means aside from obvious plays like Nvidia and Japanese robot manufacturers, it’s anyone’s guess.
Concentration, Company Size, and Fees
There are a few more differences worth noting that your average retail investor probably doesn’t care much about. IRBO is more concentrated with its top-ten holdings making up 21.36% of the portfolio while ROBO’s ten largest names account for 15.8% of their portfolio. The different approach of the two fund providers has also resulted in more large caps for IRBO when compared to ROBO. You would expect that when IRBO includes conglomerates like Lenovo, Amazon, Sony, and Hyundai Heavy Industries. ROBO, on the other hand, has less of these mammoth companies in its portfolio, and is geared towards mid-cap representation:
Geographic allocation is focused on the US in both cases, with IRBO the more concentrated of the two:
Lastly, let’s get down to brass tacks and talk about fees. This is where iShares manages to pull ahead with a total expense ratio of 0.47%, a reasonable price for a specialized thematic ETF. That’s no surprise when you consider they use a rules-based approach. Then there’s ROBO which charges just about double the fees at 0.95%. That team of PhDs they have over at ROBO Global doesn’t come cheap. It’s too early days to start looking at which fund performs better, and one could argue that until a decade has passed it doesn’t matter. We’re all in it for the long haul, right?
If we had to pick one of these two ETFs, we’d go for ROBO because they seem to be more focused on robotics and we’re not overly convinced that a rules-based approach is the way to go here. With that said, the case for investing in robotics using a rules-based approach seems much clearer than “investing in AI” using a rules-based approach. (We talked about this problem before in our article on Artificial Intelligence and Industry Classifications.) We’ve already concluded that AI is the new electricity and every company out there stands to benefit from using it. It’s those companies who aren’t using AI in your portfolio that you need to be most concerned about at the moment.
Investing in the world's most expensive paintings will provide protection during recessions. Why? Because fine art isn’t correlated to other more popular asset classes like stocks or bonds. Now, anyone can invest in paintings from names like Picasso, Monet, Renoir, Degas, Matisse, and Warhol. They're lining up, but Nanalyze readers can jump to the front of the queue here.