An asset is said to be “liquid” if traders can convert it quickly to cash without materially affecting its market price. The market for large cap stocks is liquid because equity claims are relatively homogeneous and there are normally large numbers of buyers and sellers trading on centralized exchanges. Most bond markets are highly illiquid, primarily because bonds are highly idiosyncratic. Even bonds issued by the same entity normally differ along several dimensions, including maturity, coupon rate, and covenants. Because this is so, bonds typically trade over-the-counter (OTC)—that is, in a decentralized trading environment where idiosyncratic bonds must be matched with willing buyers. These markets are typically very thin, and most bonds do not even trade on secondary markets. The bonds issued by sovereigns and large corporations are an exception. But even these bonds trade largely in decentralized OTC markets.
There has been a growing concern as of late that liquidity conditions in even relatively liquid bond markets have deteriorated in recent years. If this is so, then even modest events may trigger an unexpected and undesirable disruption in financial markets. In the summer of 2013, for example, when Fed chair Ben Bernanke hinted at a possible slowdown in the pace of Fed bond purchases, the bond market reacted violently in what was described as a “taper tantrum.” Another example is the Oct. 15, 2014, “flash rally” in which the 10-year on-the-run U.S. Treasury experienced an incredible 40 basis point movement in a single day for no apparent reason. According to a report released by the U.S. Treasury Department, it seems that for a brief period of time there were far more trades to buy Treasuries than trades to sell. That this happened in the most liquid of all bond markets raises a concern with other less-liquid bond markets. Might a modest increase in the Fed policy rate induce a “rush to the exits,” forcing a fire sale of bonds into an illiquid market to meet redemption payments?
It is, in fact, very difficult to know whether liquidity conditions are deteriorating in bond markets. Standard measures, such as bid-ask spreads, are of little help because historically narrow bid-ask spreads can widen suddenly in a liquidity event. Some commentators have pointed to the post-financial-crisis behavior of the 22 primary dealers, who play an important role as market makers for bonds. In fact, primary dealer inventories in corporate bonds have declined from over $250 billion in 2007 to about $50 billion in 2015. Since 2007, the supply of U.S. corporate bonds has increased from about $3.2 trillion to almost $5.0 trillion (see graph above), so that the dealer inventory relative to outstanding debt has dropped precipitously. Moreover, prior to 2007, dealers were net long in corporate bonds and net short in U.S. Treasuries. Dealers are now net long in Treasuries. This, together with their reduced holdings of corporate securities, suggests that dealers’ willingness and/or ability to take on risk has diminished greatly since 2008. Many commentators blame the Volcker rule, which was designed to curtail the proprietary trading activities of dealer banks.
Whether these behaviors contribute to reduce bond market liquidity is difficult to judge. In fact, one could make the case that the dealer banks are in much better position than they were in 2007 to absorb a liquidity event, for example, by absorbing a sell-off in corporate bonds with sales of Treasuries. Much of the bond supply is intermediated through money market mutual funds. Historically, these funds have sought to maintain fixed exchange rate regimes subject to speculative attack. The vulnerability of these funds to mass redemption events, however, may be curtailed with the passing of Rule 2a-7 by the U.S. Securities and Exchange Commission. This new rule requires that funds adopt a floating exchange rate regime (floating net asset value) and permits the imposition of liquidity fees and redemption gates at the discretion of the funds’ board of directors. These rules are consistent with the ones prescribed in Diamond and Dybvig (1983) for the prevention of bank runs.
It is important to understand, however, that such measures are not a guarantee against price volatility. They are simply measures to mitigate the “excess” price volatility that accompanies thinly traded markets. If everyone wants to sell bonds, their price will decline even in the most liquid of bond markets.
How this graph was created: Search for the (quarterly) series shown above and add it to the graph. Restrict the sample period to start in 2001.
Suggested by David Andolfatto