There’s an oft-quoted number in the debate raging over liquidity in the bond market.*
It is, depending on the week, 75-78 per cent — the amount by which dealer banks’ inventories of corporate bonds are said to have declined since their peak of $235bn in 2007, according to Federal Reserve data.
The Fed, perhaps in an effort to better examine the inventory issue, began breaking out the data by security type in the spring of this year.
You can see the break-down in the nifty chart from MarketAxess below. High-yield and investment grade corporate bonds are the purple and green towards the bottom of the chart. Everything else is non-corporate bond securities.
It’s a very different picture to the one you would get simply from looking at the top-line. But it doesn’t solve the problem of that 75-78 per cent figure, since the Fed never corrected the older series of data. Instead, you have to rely on analysis of obscure SEC filings (the X-17-A-5, in fact) to try to reverse engineer the full trend. On that basis, Goldman Sachs analysts reckon that dealer-banks holdings of corporate bonds are down 40 per cent from a 2006 peak of about $38bn.
Here’s the reverse-engineered Goldman chart:
And a nice one broken down by individual dealer, per those obscure SEC filings:
What to make of the revised data?
A 40 per cent drop in bond inventories is notably less than the 75-78 per cent drop implied by the Fed’s old data. Does it, as regulators suggest, lessen the banks’ argument that their inability to hold corporate bonds on their balance sheets has fundamentally changed the nature and liquidity of the fixed income world?
Here are the Goldman analysts:
Much attention has been focused on the possibility that regulatory changes like the Volcker rule and increased capital requirements have raised the costs of making markets in corporate bonds. Thus, we also note the movements in our series since the passage of the Dodd-Frank Act in 2010. Our short, gross, and net position measures declined by 27%, 21%, and 13%, respectively, between year-end 2010 and year-end 2012. Although these movements look inconsequential when compared to the dramatic swings in the mortgage contaminated Fed series over the last decade, they are still substantial changes in absolute terms. When further considering the robust level of bond issuance in recent years, these data are at least consistent with the argument that regulatory changes have reduced dealer holdings of corporate bonds.
Therein, we would argue, lies the rub.
Banks’ inventories of bonds have been declining at exactly the same time that investors, from big asset managers to ETFs, have poured money in. That is not necessarily an accident of new regulation, but a potential design feature. Investors hold more risk on their balance sheets; banks hold less. Of course, the trade-off is that investors will bear the brunt of any losses, which may be exacerbated by the lack of a bank pressure valve (so to speak), in the event of a large-scale sell-off.
* Here’s a footnote, in tribute to Matt Levine, who has left Dealbreaker and gone a-bloggin’ at Bloomberg. There is a fundamental question to be asked about the value of liquidity in the market anyway. Having a vibrant secondary market where corporate bonds change hands regularly is a great and noble thing. But it is likely to only help at the margin if and when the great bull run in corporate debt finally comes to its inglorious end. Dealers, even if they aren’t restrained by the Volcker ban on prop-trading, seem unlikely to swallow all of the bonds investors might wish to offload in heavy one-way selling.