These are the two statements you most often hear about liquidity and ETFs:

1. ETFs are only as liquid as what they own.
2. ETFs are only as liquid as what you see on the screen.

Both of these are fundamentally flawed, and interact in interesting ways. And nowhere are there more histrionics about these issues than junk bonds and their ilk (say, bank loans).

The hand-wringing seems to be coming back in vogue, as articles start popping up about the looming crisis in corporate debt (say, these comments from Greg Lippmann at LibreMax) or Scott Minerd from Guggenheim (who’s both extremely bearish and a lot smarter than I am) suggesting at the Milken conference this week that everyone pull their money from bank loan ETFs because of liquidity issues.

First, let’s get one thing out of the way — the bond market’s pretty darned big.

All told, the current outstanding U.S. debt is just about $40 trillion — or $10 trillion more than the entire U.S. equity market.

The red bars at the bottom there are corporates. Importantly, the percentage of outstanding debt that is corporate just doesn’t vary much. In 1980, 21 percent of outstanding debt was corporate. Today 21 percent is corporate. At the peak, in 1981, it hit 22 percent. At its lowest — when prices got destroyed in the financial crisis, and the Treasury issued enormous volumes — it slid as low as 18 percent. None of that says whether any of this debt is good or bad, or cheap or expensive. It’s just what it is.

More from ETF.com:

World of difference in these int’l ETFs

Stocks & rates have room to run

Vanguard begins sector transition

So what’s the truth here? Like most things, it’s complicated.

And while the daily trading was having a hard time keeping up, the reality of the past 15 years is that, for the most part, daily trading in corporates has stabilized at around 0.35 percent of the outstanding trading on a given day.

Of course, there are huge “haves” and “have nots” in there. CVS bonds may trade like water, but there are huge swaths of the market that might not trade at all on a given day. And therein lies the rub: how you think about what happens to big baskets of bonds on a day when everyone wants to sell.

My old partner Matt Hougan used to talk about how ETFs in junk bonds provided “transcendent liquidity.” I made fun of him for it at the time, but the idea is simple: An ETF like the iShares iBoxx USD High Yield Corporate Bond ETF (HYG) trades an average of $1.2 billion a day, yet the entire corporate bond market only trades $30 billion a day. That’s a billion dollars going back and forth, at miniscule spreads, while the actual underlying bonds are often barely trading.

In fact, the average bond in HYG trades, on average, only $2.5 million a day, and the least liquid bonds literally haven’t traded in the past month. Taken as a portfolio, the total volume of all the bonds on an average day is just $2.5 billion. Whether that makes HYG’s $1.2 billion “transcendent” I’ll leave to the poets.

So let’s imagine what happens on the terrible day in the future when everyone decides to get out. Imagine there’s some rash of default announcements at noon on a Tuesday. How does that information get processed by the market?

There’s actually a great theoretical framework for this in an academic article (Xu & Yin, International Review of Finance, March 2017) from last year, but I’ll summarize the basics here:

1. At noon, the ETF and the bonds are trading at the same price, call it $20. This is the old, stupid price, before we have the new information that things are awful. This is the equilibrium state.

2. At 12:01, there’s new information that the market processes. $20 is now the “dumb” price, and there’s a new theoretical “smart” price. Let’s say that’s $15. A 25 percent down day is a rout by any measure.

3. Market participants, who are always profit-motivated, will sell (and sell short!) at any price over the new, smart price. So they look for the opportunities, and lo and behold, here’s all this liquidity available in the ETF! Inevitably, a huge amount of trading happens immediately, driving down the price of the ETF. At 12:02, the ETF is now trading closer to the “smart” price, call it $15.

Note that because the NAV of the fund — in this case, the intraday net asset value — hasn’t moved yet, because in this moment, none of the underlying bonds have changed prices, and no bond services have issued new indicative prices for the less liquid holdings. So we’d now say it’s trading at a discount.

4. Now it gets interesting. At 12:02, there’s now this big discount happening, and a new set of market participants enter the fray — arbitrageurs. In this case, we have authorized participants, who know that they can swap the actual bonds for shares of the ETF, and vice versa. They now come in and do two things.

They start buying up shares of the ETF, while simultaneously selling the bonds, putting real prices in the market. This pushes the ETF trading price and the value of the bonds closer together, and slows the decline of the ETF in the process. At this point, the less liquid bond market comes online in full force, and the selling pressure on the ETF’s holdings will begin in earnest.

5. By the end of the day, these counteracting forces and disparate market participants have had time to absorb the new information, and work out any friction from the system, and the ETF and the net asset value settle back into equilibrium at the new, smart price of $15.

It’s also worth noting that, while the above straw man posits a sudden shock, we see this play out every day in ETFs. When lots of money wants in, ETFs tracking less liquid securities like junk bonds trade at a perceived premium until the market shakes out the noise. When lots of money wants out, the opposite happens:

The problem — particularly for folks who aren’t living this stuff day to day — is in how you perceive what’s going on here. Without context, you can spin the (false) narrative that “the ETF went down, and crashed the bond market.” In fact, what happened is that the market — the whole market — decided that bonds were overvalued; thus, prices were going to come down no matter what.

All the ETF did was provide for a more orderly mechanism of implementing the price discovery. (For folks who want to wade more into those deep waters, I’d point you to several excellent pieces from BlackRock on this topic, here and here).

But it’s important to note — if, at the end of the day, people really hate the underlying assets here, and there’s really nobody who wants to step in and be the buyer of last resort, there’s nothing magical about the ETF that will provide a bottom.

In the above hypothetical example, junk bonds are junk bonds, and if everyone takes their marbles and goes home, well, they’re going to crash. The opposite is true of course — if everyone starts piling into junk bonds in a hurry, they’ll “crash up” just as fast.

The lesson here for investors is one I repeat so often I’m thinking of getting a tattoo: Know what you own and why you own it.

If you’re buying into a volatile, illiquid corner of the market, you should be prepared for how your investment will react to new information. The fact that you can get in and out through a highly liquid ETF is a convenience, but it’s not a panacea that makes risk go away—it just makes the process of engaging and disengaging that risk easier.

— By Dave Nadig, managing director of ETF.com

Dave Nadig can be reached at dnadig@etf.com