Nasdaq’s Phil Mackintosh recently examined a single chart that combines many of the things he’s recently been discussing regarding market fragmentation, routing, spreads and ETF liquidity.

What he found is just how differently popular ETFs trade versus thinly-traded ETFs.

Smaller ETFs trade differently than large ETFs

In Chart 1 (see below), each circle represents an ETF ticker. Note that the biggest ETFs (big circles) also see the most trading (in the bottom right corner). That in turn results in the most competitive, and tightest spreads (low on the vertical axis). So far no surprises for regular readers of our posts.

However, where this chart gets interesting is the color scheme. Yellow circles represent stocks that trade on-exchange more than the market wide average. While the darker the circles get, the higher the proportion of trading that occurs off-exchange.

What that highlights is that for many thinly traded ETFs, the proportion of trading off-exchange is very high. There are a couple of good reasons for this.

First, the underlying investors are different. Hyper-liquid ETFs are frequently used by hedge funds and banks to hedge and take tactical positions. That’s why their turnover often looks more like a future than a stock. This in turn creates economies of scale for market-makers and arbitrageurs, even leading to tighter quotes and more tactical trading.

In fact, their spreads become so tight that they are usually much cheaper to trade than the underlying stocks.

In contrast, many of the smaller ETFs are designed for longer-term investors with active or smart beta portfolios. That in turn makes them more popular with retail and investment advisors. As we noted recently, advisors and retail investors tend to trade much more off-exchange.

We’d prefer to see tighter spreads on exchange and made some proposals to deliver this in our ETF Rule Comment. But until that happens, off-exchange trading of thinly traded ETFs can often help reduce trading costs, as we discuss below.

Chart 1: As ETF popularity increases, spreads fall and exchange trading increases

Source: Nasdaq Economic Research

Don’t be scared by ETFs with wider spreads and low ADVT

One of the unique things about ETFs is that market-makers can buy the ETF or buy the basket and do a creation. That’s important because it means, for larger trades:

ETF daily volume is NOT important, as arbitrageurs and market makers can pull liquidity from the underlying stocks into any ETF. That means most ETFs can easily absorb trades of at least $100m, which (for perspective) is roughly the market-wide value on offer in the U.S. market at any time during the day
ETF spreads should NOT deter investors from considering larger trades in an ETF, as market makers can buy the underlying stocks and convert that cost to an ETF
However, by definition, when you trade through a market maker you contribute to increased “off-exchange” trading.

Even if your trade is small, crossing the ETF spread is usually still cheaper than buying the basket of underlying stocks. In fact our data shows that more than 99.4% of ETF liquidity (by U.S. dollar value) trades with spreads less than 20bps.

Bonus Chart: What about Stocks?

In case you’re wondering what this looks like for stocks, we’ve added that chart below too.

What we see is interesting: stocks trade more “the same” with many stocks being a pale shade of yellow (slightly more lit trading than the market-wide average). Although we do still see spreads are tighter as daily increases (diagonal slope of the line).

Thinly traded stocks (left side of chart) do tend to trade more dark, the pattern is nowhere near as strong as for ETFs—nor does it start at the same point:

ETFs trading less than $20 million/day tend to commonly trade a high percentage off-exchange
While stocks trading less than $1 million/day sometimes trade a little more off-exchange than average
One factor that is likely contributing to both stocks and ETFs is that retail internalization is a higher proportion of trading in smaller-cap securities.

Chart 2: The same chart looks very difference for corporate securities

More from Economic Research

Three Charts that dispel the Price Improvement Myth

The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.

Refinitiv announced new research findings that confirm the use of machine learning is pervasive across the financial community and is critical to its success in the future. Ninety percent of the c-level executives and data scientists surveyed have already deployed machine learning, while all of the c-level participants said it is core to their business strategy.

However, respondents acknowledge that poor-quality data impedes their ability to fully leverage machine learning and artificial-intelligence technology, with 43% citing this as the biggest barrier to adoption followed by a lack of data availability (38%). Despite being a technology area that’s seen a recent ‘war for talent,’ challenges around data quality were ranked ahead of access to talent, which was highlighted by a third of the respondents.

This information was compiled for the inaugural Artificial Intelligence/Machine Learning (AI/ML) Survey by Refinitiv, featuring in-depth interviews of nearly 450 financial professionals across North America, Europe and Asia. Its findings confirm how far the industry has evolved since 2017 research that indicated technology companies were the primary adopters of artificial intelligence (AI) and only 28% of financial-services firms were deploying it. [1]

Key findings from Refinitiv’s new research include:

90% of financial firms are using machine learning, either in multiple areas as a core part of their business (46%) or in pockets (44%); the 10% of firms that have not yet deployed machine learning are experimenting with it
75% of firms are making significant investments in machine learning
62% of c-suite respondents plan to hire more data scientists in the future as banks and asset managers seek to give themselves a data and technology edge over competitors
The main applications for using machine learning were in risk management (82% of respondents), followed by performance analytics and reporting (74%), with alpha generation in third place (63%)
AI/ML adoption is primarily driven by extracting better quality information (60%), increased productivity and speed (48%), and cost reduction (46%)
“Machine learning and artificial intelligence are often described as emerging technologies, but the fact is they are already being widely applied across financial services,” said Tim Baker, global head of Applied Innovation at Refinitiv. “Whether it is an increasingly complex regulatory environment, the need to find new sources of alpha, or winning the fight against financial crime, the industry is turning to data and technology, and data scientists are increasingly important as the alchemists charged with turning big data into insight.

“We see a future of accelerating innovation fueled by wider availability of powerful cloud-based artificial intelligence and machine learning tools dramatically lowering entry barriers and thus changing the competitive dynamic across the industry. But no financial institution will be able to use the technology successfully unless the underlying data is machine ready.”

Learn more about the Refinitiv AI/ML survey.

One thing for sure, it isn’t just stoners like Cheech and Chong any more. And these new users and those non-users embracing legal cannabis are driving the market are spread wide and far across the demographic spectrum, according to new research from Quinn Thomas and DHM.

As noted in a recent commentary from DataTrek, co-founder Jessica Rabe thought it was important to examine just who is driving the push for cannabis legalization and dispel notions that marijuana isn’t just for movie stoners or older hippies.

The Quinn Thomas/DHM research sampled 900 marijuana users in Colorado, Washington and Oregon and also conducted two focus groups in Portland and Seattle. They chose those three states because Colorado was the first to allow recreational sales over five years ago, followed by Washington and Oregon, so they have the longest track records. Their research also notes that “Oregon has the highest percentage of monthly cannabis consumption in the country among its adult population – 20%”.

Let’s take a look.

Cannabis User Demographics:

Gender: Male (60%), Female (40%)
Age: 22-34 (30%), 35-54 (+35%), 55+ (35%)
Household Income: Less than $25k (24%), $25-50k (23%), $50-75k (19%), $75-100k (14%), $100-150k (14%), Over $150k (6%)
Relationship Status: Single (33%), Married (51%), Cohabiting with long-term partner (11%)
Party: Democrat (35%), Republican (26%), Independent (32%), Other (7%)
Education: Less than high school (30%), Some college/2-year degree (33%), College (19%), Graduate degree (16%)
Rent or Own: Rent (32%), Own (60%), Something else (8%)
“It is a population that has much more in common with a typical middle-class resident than with the caricatures created in Hollywood movies,” noted DataTrek’s Rabe.

Users generally match the U.S. average educational attainment, household income levels, and the political composition and race and ethnicity breakdown of the three states researched, noted the survey. This includes everyone from young-single renters to middle-age homeowners with kids to retired seniors.

Rabe added that there were two caveats to this: First, men consume marijuana more than women. Secondly, frequent users cluster at the lower end of the income scale. For example, 35% of regular users (who consume almost daily) have household gross incomes of under $25k per year compared to 11% of infrequent users.

So, do users drink along with their smoke?

The survey found that 38% of respondents drink less when they canna-indulge, 58% drink the same amount and only 4% drink more. And of this group, 24% said that they replace alcoholic beverages with cannabis.

“Over one-third of marijuana users drinking less post-legalization is a significant stat, especially since that’s across the board in terms of frequency of use,” Rabe noted. “Of course, it’s truest for regular users (near daily) with 58% drinking less since legalization. But 32% of occasional cannabis users (7 to 8 times a month) and 26% of infrequent marijuana users (2 to 3 times a month) report drinking less as well.”

DataTrek also pointed to some other statistics from the Quinn Thomas survey

Consumer Behavior

Comparing cannabis use before and after legalization:
Regular consumers who consume daily: Pre-legalization (48%), Post-Legalization (74%)

Occasional consumers who consume a few times a month: Pre-legalization (11%), Post-legalization (27%)

In what situations do you typically use cannabis?
On weekdays, during the work week: Yes, often (29%), Yes, occasionally (34%), No (37%)

On weekends and holidays: Yes, often (39%), Yes, occasionally (46%), No (16%)

By yourself: Yes, often (39%), Yes, occasionally (35%), No (27%)

So, what does Rabe come away with?

First, she said recreational marijuana legalization has led to greater demand across the board, from regular users to even infrequent users.

“If you’re invested in public marijuana companies, or want to do so, this is clearly a positive for the industry’s growth rates. That said, one blind spot in the report was its lack of mentioning the black market. We keep emphasizing that the legal cannabis industry’s total addressable market hinges on appropriate regulations and reasonable tax rates. Marijuana taxes in California and Washington, for example, are too high and support illegal sales. Since the black market has existed for decades, it’s already entrenched and efficient across the US. Regular users, in particular, likely already have “a guy” and won’t want to incur high tax fees even if they can buy it from legal dispensaries.”

Secondly, the more cannabis people consume, the less they drink.

“We also continue to note this important substitution effect because it has a meaningful impact on public liquor companies. Many have already been vocal about this risk. Liquor and beer sales are slowing or declining and further marijuana legalization across the US will likely accelerate these trends. That’s why liquor companies, such as Constellation Brands and Molson Coors, have either taken a stake in a Canadian marijuana company or have partnered with one to make non-alcoholic, THC-infused drinks.”

She added that if one is invested in the liquor industry, it’s important to understand that large public companies in that space haven’t just taken notice, but are already taking action to diversify their product offerings or business models. Those who don’t will likely lag behind.

And lastly, the report noted that their focus groups associated alcohol with “going out” and marijuana with “staying home”. Rabe said this is likely because consuming marijuana in most public places remains and is illegal.

“We think this will change as laws loosen, the cannabis stigma fades, and more legalization rolls out across the US,” she said. “The industry has already started developing an entire new distribution channel similar to bars and lounges. Consumers can go to marijuana cafes in Washington, for example. It’s still early days, but marijuana is creating a platform where eventually consumers don’t have to go to a bar for liquor, but can go to a cannabis club, marijuana bar, or of course just consume the drug at home.”

To see the entire Survey report from Quinn Thomas/DHM, please click here: https://www.quinnthomas.com/wp-content/uploads/2019/03/cannabis_next_door_report_march2019.pdf