It turns out that a booming Wall Street business is based on guesstimates, and that could be a problem
You may have heard that people are worried about bond-market liquidity.
Wall Street executives complain about the dearth of liquidity so often that Bloomberg's Matt Levine has a section in his daily email dedicated to detailing their gripes .
Concern about "liquidity" means that when fund managers want to sell in size and at speed, the market won't be there to absorb the new supply, sending prices spiraling.
In a big note out Friday, Oleg Melentyev and Daniel Sorid of Deutsche Bank took a look at what they called the "liquidity vacuum," and we have to say that the report is full of really interesting stuff.
The big picture is that liquidity has deteriorated since they last took an in-depth look a year ago. The market, however, is adapting.
From the note (emphasis ours):
If dealer inventories were a concern back then, when they were running $5bn in HY and $13bn in IG, they must have become even more so by now. Dealer community is barely averaging $1bn in HY inventories, less than 15% of daily turnover in this market, and $4bn in IG or 20% of turnover. In fact, a look inside these headline numbers suggests that dealers are routinely short the largest segment of both markets — 5-10yr maturities — while compensating with excess longs in short durations.
So dealers have basically left the scene of principal trading in corporate credit. Practically speaking, the market is now operating almost fully on an agency basis, with implications being lower daily turnovers, particularly in IG, wider bid-ask spreads, higher incidents of one-sided markets. The market is nevertheless learning how to operate in absence of principal dealer bid, and it does so with mixed success, sometimes gapping around turning points, but generally finding a way to price the transfer of risk.
The reduced liquidity is really most striking for bonds that have been around for any length of time. Basically, bonds see a lot of trading right after they are issued, and this trading deteriorates sharply over time.
Here is Deutsche Bank: "Between days one and five, average volume drops by 80% in HY, and by 65% in IG, at which point it continues to decelerate going forward, but at a much slower pace. Following the sharp drop in the first week, It takes another 20 days for an average HY and IG bond to lose another 50% of its volume."
Now, that loss of volume is important.
Bonds that are one year old make up 20% of the benchmark but 35% of trading activity. At the other end of the spectrum, you have bonds that also make up a chunk of the benchmark but barely ever trade.
These benchmarks that include bonds that do trade, and thus have pretty transparent market-based prices, and bonds that don't trade a lot, which lack fresh pricing information, are used to create things like exchange-traded funds.
Credit ETFs have exploded in popularity. The daily turnover in the equity exchange-traded shares of the two largest high-yield ETFs (HYG and JNK) for example now stands at about $1.75 billion, double what it was in mid-2015. Investors seem to be turning to these ETFs to get exposure to the credit market quickly and easily and then replacing the ETFs over time with the harder-to-buy bonds.
The result is that ETFs, which offer daily liquidity, are having to guess at prices for the bonds that aren't trading. And those guesses could throw out a whole bunch of other calculations.
Here's Deutsche Bank on why that a problem (emphasis ours):
So there you have it: roughly one-third of all names in HY and IG barely ever trade. They could go for months without a real price print, meaning that all those indexes tracking thousands of ISINs on a daily basis are routinely guessing values for 25-40% of their constituents.
It is what it is, whether we like it or not, the reality of the situation, and at least to this point, the industry has not come up with a better solution to this problem than guesstimating values for a large number of bonds. But it nevertheless provides an important lesson to all those using indexes to measure their volatility, sharpe ratios and other useful risk/return characteristics. Because a large portion of this indexes is marked at stale prices, all these otherwise useful measures are routinely underestimating true volatility in credit. This in turn could lead to mis-valuation as arguably investors are looking at their expected risk-return profiles to make allocation decisions. If a bad volatility estimate enters that analysis, its conclusions are bound to be flawed. Investors would be better off using volatility estimates derived from real traceable credit assets, such as CDXs, ETFs, or baskets of liquid bonds.