Over the past few months there's been a lot written about a shadow hanging over bond markets: the shadow of a liquidity squeeze.
The latest to wade into this debate is Maya Bhandari, director of multi asset allocation at the U.S.'s 11th largest asset management firm, Columbia Threadneedle. She looked closely at the doomsayers' case and found it wanting.
According to their thesis, with the prospect of interest rate hikes looming, the great bond bull market may soon be dead. The fear is that, with the global reach for yield having squeezed the potential of further gains from bonds and large players exiting the intermediary market, price moves in response to rate rises could be violent.
In this scenario, the liquidity that traders thought was there suddenly evaporates at the point where they need it most.
Bhandari said that while turnover and block trade volumes have fallen in recent years, there's little solid evidence of a regulatory squeeze on trading capital.
"Bond markets have not deviated from their fitted curves, as they would if liquidity indeed were more costly (as the costs of arbitrage to eliminate those divergences would be higher)," she said. "Nor, crucially, is it clear that banks and other intermediaries would behave in a counter-cyclical manner anyway and provide liquidity in tougher market conditions. Dealers and brokers, and indeed, investment banks, appear to be strongly pro-cyclical in their behavior, with a firmly positive relationship between changes in leverage and changes in bank sheet size."
Bhandari's comments echo the findings of Hyun Song Shin and Tobias Adrian in their 2010 paper Leverage and liquidity, which showed the size of dealers' balance sheets appear to have a strong procyclical bias. That is, they tend to increase the size of their balance sheets during booms and shrink them during busts due to changes in the calculation of value-at-risk.
Shin and Adrian argue that balance sheet size determines aggregate liquidity and, given that balance sheets expand through increased collateralized lending and borrowing by financial intermediaries in booms, the unwinding of that leverage during busts means it's extremely difficult for primary dealers to act as buyers of last resort.
"Procyclical leverage then translates directly to counter-cyclical nature of unit value-at-risk (i.e. value-at-risk per dollar of assets)," they wrote. "Measured risk is low during booms and high during busts."
As such, to the extent that higher capital requirements imposed by regulators limit both the balance sheet leverage of these institutions and the inventory capacity of primary dealers during booms it is difficult to see why, at least in theory, asset price movements would be more volatile over the economic cycle. If anything, the reverse might be expected to be true.
Anyone arguing that current conditions in bond markets are clearly signalling a liquidity squeeze have to explain why these concerns are yet to show up in key indicators of stress.
In its October series on bond market liquidity, the New York Fed found that although dealers have shifted from a principal model to an agency model, whereby they match orders without having to hold bonds on their balance sheet, bid-ask spreads in corporate bond markets are currently even lower than their pre-crisis levels. The authors take this as evidence of "ample liquidity" in these markets.
To believe a crash is inevitable, it's necessary to believe either corporate bond markets are oblivious to the problem and are systematically mispricing liquidity risk or that market participants are seeing the pitfalls but don't believe that there is much of an issue.
Part of the apparent contradiction between falling dealer inventories and otherwise benign looking conditions in bond markets could be explained by the fallout from the financial crisis.
"If you exclude non-agency mortgage-backed securities you see inventory levels have remained roughly the same since around 2002," Bhandari said. "Much of the additional liquidity you saw in the intervening years was the result of an unhealthy build-up of MBS in the run up to the global financial crisis."
Importantly, she said that the modelling work done by Columbia Threadneedle suggests that the theoretical and empirical gauges of liquidity are broadly aligned, meaning investors' perception of what they ought to be paid to take the risk of holding bonds approximately matches the compensation they are receiving from the market place.
However, there are some reasons to worry that investors could still be mispricing credit risk.
With the prolonged period of low interest rates and large scale asset purchases programs having flattened the yield curve, it is possible that investors could be caught out once central banks begin to tighten policy - especially if inflation rises faster than markets expect. This is likely to be exacerbated as the limited upside in investment grade and government bonds mean that they become much more correlated with the performance of risk assets such as equities than they have in the past.
Equally, with default rates in U.S. energy stocks ticking up it is possible that some of the reach for yield in recent years could come back to bite some investors who have significantly increased their exposure to speculative grade credit in the sector.
A rush to the exit triggered by either of these scenarios could be unpleasant for bond investors.
Yet these concerns have little to do with a liquidity crunch caused by regulatory pressures. While it is clearly in the interest of market makers to complain about regulatory overreach, the evidence that this poses a systemic risk remains limited.