Exactly two months ago, Goldman joined the relatively small group of voices warning that the next big market "event" (politically correct way of saying crash) could be one resulting not from excess leverage but lack of liquidity, driven by "phantom liquidity" provided by HFTs and algo traders which tend to withdraw during times of market stress, as well as central banks which have soaked up a substantial amount of the freely-traded securities in the market.
Specifically Goldman warned that "the rising frequency of “flash crashes” across many major markets may be an important early warning sign that something is not quite right with the current state of trading liquidity."
These warning signs plus the rapid growth of high-frequency trading (HFT) and its near-total dominance in many of the largest and most widely traded markets prompt us to more carefully consider the possibility (not necessarily the probability) that the long expansion accompanied by relatively low market volatility may have helped disguise an under-appreciated rise in “market fragility.”
To be sure, the topic of rising market fragility is anything but new to regular readers, and we have been covering it extensively over the past two years, although Goldman's growing concern by what was painfully obvious to many traders gives hope that one day the Fed too may be able to grasp just how its actions have broken the market, although that realization will sadly take place just moments before a market-wide flash crash send the S&P plummeting, resulting in a market that could be indefinitely halted.
Fast forward to today, when Morgan Stanley has joined the chorus of "liquidity watchers."
In a note from Morgan Stanley strategist Andrew Sheets, he writes that whereas 2017 was remarkable - "despite low
levels of volatility that made the bar for a large move relatively low, few occurred" - 2018 has been very different in that "large moves relative to expectations are becoming more common" even though it may - still - not feel that way, with volatility still low and US equities back near local highs. "And that's exactly why it's so interesting."
First, what does Sheets means by by a 'large move'? He defines it as a one-day change that is a 3-sigma event or greater, relative to what was implied by options markets at the time, in the context of 3 specific reasons:
Expectations: By comparing moves relative to what was 'implied' by options, we obtain a direct measure of the move relative to expectations. And we think it is relative to expectations that moves really matter. A 1% move when the market was expecting daily gyrations of 0.5% is a surprise. That same 1% move when the market is expecting 2% is a relief.
Scaling: Since 2000, the average daily move for oil prices is 1.5%. The average daily move for EURUSD is 0.4%. Scaling clearly matters for what should qualify as 'large', and adjusting by volatility does this.
Subjectivity: 'Large move' is a subjective term. By counting only 3σ events, we can arrive at a consistent definition of 'large' across time and asset classes. Different thresholds are an easy adjustment.
So after looking at 14 different asset classes across four key groups – equities, rates, FX and commodities - and tabulating the number of cases where a one-day price change would have been a 3σ move based on the prior day's implied volatility, here is what Morgan Stanley found:
2018 has seen a meaningful uptick in large moves relative to option-implied expectations across most asset classes. The contrast with previous years is most pronounced in global equities, which are on pace to see the highest number of such moves since 2008.
When the bank aggregated the moves across asset classes, the trend was clear: 'Large moves' are becoming more common in 2018, and are running at the highest rate since 2008.
Performing the same analysis for "just" 2-sigma threshold moves reveals a similar uptick in unexpectedly "large moves" across asset classes.
Now, to those who have read our frequent posts on how HFTs and central banks have broken the market, Morgan Stanley's findings should come as no surprise, and yet they represent the latest warning from a top financial institution that something is badly broken with the market, a warning which follows similar caution from Goldman, which we discussed in mid June, and in which the bank's co-head of global equity trading Brian Levine made the following disturbing preview of what his own personal trading hell looks like, one in which the market "flash crashes" but then fails to rebound:
The data is wrong, everything trades at dislocated prices relative to the NBBO, and everyone—justifiably—widens their spreads. That happens almost every time there’s volatility, largely because message traffic increases dramatically. This is due to the fact that the opportunity set is greater and there’s no economic disincentive for sending messages to the market, so more electronic orders come in. This slows the system, widening spreads and generating price dislocations, which triggers even more orders and compounds the delays—a predicament that is only further exacerbated by the fragmentation of the equity markets. As this happens, stocks may trade outside of the NBBO briefly in millisecond or microsecond increments, constituting what I consider a genuine flash crash. All of this becomes a negative feedback loop that causes more volatility.
Interestingly, if you define a flash crash by the percentage of executions that took place outside the NBBO, one of the largest ones occurred in 2008 after the first TARP bill failed, according to internal analysis we did a few years ago. And the market didn’t snap back, with the SPX closing down 10% on the day and on its lows. I think that may have been why there wasn’t talk of a “flash crash” afterward, but clearly the market structurally failed pretty badly that day, too. This suggests to me that, in a situation with actual bad news, the current US market structure may not be able to handle it, and there could be a downward spiral.
And while Goldman blames HFTs and algo trading as the culprit behind the market's increasingly more frequent flash crashes and "breaks" (while occasionally pointing the finger at central banks as well), Morgan Stanley has a different set of explanations for the uptick in large moves. The bank notes the usual suspects, including: tightening policy, extreme sentiment towards equities and USD to start the year, trade tension and geopolitical risks, while also noting that "volatility has remained generally low in 2018, lowering the hurdle for a large move."
But the explanation the bank's strategist finds most worth discussing is liquidity "or, more accurately, the lack thereof" to wit:
The fact that constrained liquidity is present across major markets mirrors the broad-based uptick we've seen in outsized moves.
Below we list the most relevant explanations according to MS:
Markets have grown. Dealer capacity has not
The bank's argument here is simple: while asset markets have seen a substantial increase in notional, dealer capacity has failed to increase appropriately, or as MS writes, "It may not feel like it, but financial markets are significantly larger than they were a decade ago."
Consider the following, comparing July 2008 and today:
- S&P 500 market cap: US$11.5 trillion in July 2008. US$24.8 trillion today.
- EUR sovereign bond market: €4.6 trillion in July 2008. €7.5 trillion today.
- USD aggregate bond market: US$10.7 trillion in July 2008. US$20.1 trillion today.
- EM sovereign bond market (this includes EMBI-eligible sovereigns and quasi sovereigns and excludes private corporates and non-EMBI sovereigns): US$288 billion in July 2008. US$894 billion today.
Yet while markets have grown steadily over the last decade, the means to trade them have not. The last 10 years have seen a historic deleveraging of bank balance sheets globally, a response to the clearly overextended state of balance sheets prior to the crisis.
Sheets brings attention to credit markets as the most directly measurable, and stark, example of this with the chart below:
On the left-hand axis of Exhibit 10, we plot the total size of US credit markets, as proxied by the combined size of the Bloomberg Barclays IG and high yield indices. On the right axis, we plot total dealer holdings of US corporate bonds – a significantly larger market with a lot less inventory on the shelves. Dealer holdings of corporate bonds have shrunk from 3% of the market to just 0.3% today. While this means that dealers themselves have less to liquidate, their capacity to move risk to a new buyer may be limited and require larger repricing of the asset class in times of stress.
Central bank dominance
A familiar argument, one we have made frequently in the past: the longer QE went on, the more securities ended up on central bank balance sheets - including Treasuries, ETFs and equities - vividly seen in this chart of freely floating German bunds...
... and the result has been a sharp drop off in overall liquidity. However, an even greater risk is what happens to the liquidity as measured by "market depth" once the ubiquitous central bank bid goes away.
In Morgan Stanley's words, as traditional banks pulled back, central banks became significant market players, accumulating quantities of assets over the last 10 years. Central banks hold 28% of the Agency MBS market (the Fed), 22% of the European sovereign market (the ECB), ~10% of the European IG credit market (the ECB again) and ~42% of the JGB market (the BoJ).
As central banks built these positions, liquidity in the affected assets was excellent. It's hard to imagine anything better for liquidity than the presence of a steady, deep, well-telegraphed bid. But these forces are now swinging in the other direction. The Fed's purchases have already begun to reverse, the ECB's are likely to over the next six months, and with close to half of its bond market already owned by the BoJ, it will eventually face a constraint.
Finally, Morgan Stanley notes that liquidity has started to affect even futures markets, traditionally the biggest stores of wealth, as the trades do not directly affect the underlying security, and are thus immune from asset scarcity. Or at leasst were: "As reduced inventories constrained liquidity in traditional products, and central banks tilted the scale on others, more volumes have moved towards futures markets as a means to gain more liquid positions. Futures trading positioning has soared, as measured by the open interest."
Ironically, futures can find themselves a victim of their own success: if you are the most liquid asset in a stressed market, you may be more likely to be sold because you can be, rather than because you should be. Elsewhere, volumes in the FX market have been drifting lower in the last year.
The chart below shows the increasing open interest in the TY (Treasury) market, even as volumes in the multi-trilion FX market have been falling.
* * *
Taking all these observations into consideration, Andrew Sheets concludes that as a result of the higher, if still generally underappreciated, liquidity and gap risks skew should be steeper across markets: "this is especially true in illiquid assets like credit. Skew is off the lows in several markets and closer to average, but we think that it can persist at these levels and rise further."
His second conclusion goes to the very nature of how we measure market volatility, and warns that "investors should be cautious about extrapolating recent realised vol as a measure of risk and an input into position sizing."
His conclusion, while not as stark as that presented by Goldman's co-head of equity trading, is nonetheless relevant for those traders who have a substantial exposure to what has emerged as the biggest market risk: liquidity, or lack thereof:
The liquidity regime can change suddenly and using through-the-cycle measures is important this late in the cycle. The low realised volatility environment of 2017 is a poor parallel for the year ahead.
As usual, this latest market structure warning will be neglected until just a few seconds after the fact, at which point the fingerpointing will begin.