In a recent tweet I made the comment “Not a paradox: ratio of MS to assets & of asset prices to GDP, and hence to GDP functions C/I/S/P out of synch”.
Given the sensitivity of markets to even small changes in demand should anyone stand ready to provide liquidity at the onset of the tail of a distribution?
If a liquidity decision eventually exits, then there is a maximum amount of divergence which any given financial system can accommodate before the dynamics of the reverse flow overwhelm any attempt to keep it afloat. We must bear this in mind.
A recent article by Nouriel Roubini “The Liquidity Time Bomb”, to which the tweet responded, commented on the apparent paradox between vast amounts of financial stimulus and monetary expansion alongside a decline in market liquidity for assets.
Why do we need liquidity in the market place? There are a variety of fundamental economic reasons. Entities wishing to purchase assets (from savings out of income or from loans) in either new/existing issues, entities who may be dissaving and wish to sell assets in exchange for cash for either consumption (debt repayment) or to purchase higher yielding assets, businesses that wish to raise capital and other entities like governments that wish to borrow. That is markets function as an important medium for saving, consumption, investment and production decisions…they facilitate the allocation, pricing, accumulation (and the reverse) and transfer of capital ownership rights.
The relationship between these flows and their economic fundamentals should be pretty tight at an economic level where expenditure, saving, investment and production functions remain fairly stable over time: if you were to model your inflows and your outflows over time, the % annual net transaction (+ve or –ve either way) is likely to be modest and possibly matching of net dynamics in the real economy. Careful monitoring of the dynamics of flows should help pin point periods where there are imbalances between demand for and supply of assets relative to C/S/I/P functions are extreme. This is more or less a mirror image of the asset and liability modelling that portfolios should be conducting over time.
In a balanced world the provision of liquidity to finance key functions should be fairly easy to price assuming that structures designed to manage risks are themselves designed around the ability to pass through periods of extreme volatility: if the asset and the liability side of markets and economies are in a close relationship all but the most extreme events should be capable of being easily negotiated.
A point about extreme events is worth making: a distribution at equilibrium is assumed to record the response/sensitivity of the frame to random independent shocks to the system. Much of the distribution towards the tails in financial markets are not recording this dynamic but the reaction of a system either out of equilibrium or in an accommodated equilibrium to shocks, and these shocks themselves are more likely to be the result of imbalances that already lie in the system. So it would seem rationale that liquidity provision acknowledges this reality as should portfolio structure and management.
In a world of rationale long term investors where primary decisions are driven by the time frame and distribution of liability profiles and their risk preferences the boundaries of the volatility of core liquidity requirements should be fairly easily established. Likewise in a model with fairly tight parameters and fairly certain known liabilities the likely distribution of the changes in market demand is likely to be again fairly tight.
This has not always been the case due to a number of factors:
It requires only small changes in desired cash balances as a % of assets to significantly change market prices. If the market wants to reduce its cash allocation from 10% to 5% of assets and there is only 1 other asset at 90%, the only way it can do this is to increase the value of the other asset till cash as a % falls to 5%;
Apparent investor risk preferences can change markedly in short time frames;
Changes in asset focussed money supply growth, especially those that accumulate over time can alter the dynamics of valuation, impacting both A and B at key moments.
And, of course, any factors which serve to further focus the monetary and funding dynamic towards asset prices: note increasing income inequality and unusual monetary policy (QE, zero to –ve IRs).
Given that small changes in demand can significantly alter asset prices, any significant accumulated differences between the dynamics of the market relative to the dynamics of the economy should cause large changes in a market’s liquidity pricing: where accumulated excess is divergent from economic fundamentals you also have to question any rationale that says liquidity should be available, in copious amounts, to resolve its exit. Rationally speaking, excess needs to be driven out or at least dissuaded from entering a complex system of importance and exit pricing must be a factor in this decision.
Critically when we speak of liquidity we need to understand the hierarchy, timeline and transaction spirals that led to market allocations. Assets are purchased with a timeline, a timeline that is usually dependent on risk preferences and liability profiles and valuation perspectives, and it must therefore be acknowledged that these transactions are going to reverse. Where we have excess money supply growth, a search for yield and excess allocations to higher risk assets that may not match longer term C/I/S/P decisions, we are more likely to arrive at a point where the transfer requests out vastly exceed the transfer requests in. Imbalances ignore the fact that transactional liquidity is a barbed arrow: it must at some point come out and it is therefore imperative that financial market dynamics and structure are reflective of C/I/S/P dynamics to manage these risks.
At the present moment in time we have experienced a fairly long period of excess asset focussed broad money supply growth. This at a time when global economic growth has slowed and risks to growth have become increasingly elevated. Additionally we have increasing income inequality which has further increased the asset focus of money supply over and above C/I/S/P dynamics and an unusual monetary environment which has encouraged its asset focus.
So we have a situation where the future economic flows of income/revenue growth are at risk of being odds with asset values such that asset values effectively exceed the ability of future flows to facilitate the transfer of assets at current prices. We have an implicit liquidity deficit, with liquidity itself being a function of the relationship between present values and future flows.
I guess in statistical terms we have a latent and implicit price reaction that would reside in the far left corner of the tail given that demand flows for assets exceed the future flows of natural economic relationships: in other words latent exit dynamics exceed latent entry dynamics. Given the sensitivity of markets to even small changes in demand no one should stand ready to provide liquidity at the onset of the tail of a distribution.
So we come back to “Not a paradox: ratio of MS to assets & of asset prices to GDP, and hence to GDP functions C/I/S/P out of synch”.
Monetary demand for assets and the impact this has on prices relative to economic fundamentals places asset prices at a significant divergence from inter temporal flows of economic activity. Why provide liquidity for short sharp readjustments between the two? There is no rationale for providing liquidity to the outer bounds of a distribution. In a sense, unless there is a genuine C/S/I/P imperative we are all hostages to market pricing and liquidity provision.
But the concern is that the build up of excess subverts the ability of the market, at key moments and for significant time periods, to provide key liquidity to fundamental transactions. That surely must be the concern! This is a systemic risk but one only given prominence through vast financial sector imbalances, as are most of today’s systemic risk issues.
There is a fairly large body of work on market liquidity dynamics, structures and risks, some that argued that historically the financial sector charged far too much for its intermediation, others that reflect the impact of the introduction of algorithmic high frequency trading on costs and market efficiency and others that address some of the more complex issues created by collateralisation in the derivatives market as well the increasing impact of Central Bank action on liquidity in the fixed interest market place.
A large part of the problem we face today is the increasing divergence between the financial market and the economy and the complex structures that have developed as a result not only of this liquidity on asset values but the risks that arisen as a result and which have themselves been hard coded into the system as they are themselves laid off and hedged. We are in a large cinema with a few small exits and are unfortunately witness to a large number of pyrotechnic displays throughout. There is an increasing body of work that has attempted to understand the connections between financial institutions, that is the numerous blood vessels and capillaries that have arisen as the “financial body” has arisen, but such a study may only shed more light on the scale of the problem than offer any meaningful attempt at containing it. As I said before, transactions that bring liquidity in ultimately reverse and the balance between the market and C/S/I/P decisions of the economy is critical to containing risk and to ensuring that liquidity and transactional demand requirements match.
And finally, if a liquidity decision eventually exits, then there is a maximum amount of divergence which any given financial system can accommodate before the dynamics of the reverse flow overwhelm any attempt to keep it afloat. We must bear this in mind.
And some further reading:
We see signs that market liquidity is increasingly concentrating in the most liquid securities, while conditions are deteriorating in the less liquid ones (“liquidity bifurcation”). The trend can be seen in both the supply of and demand for market-making services, and reflects both post-crisis cyclical conditions (such as diminished bank risk appetite and strong bond issuance) and structural changes in the markets themselves (such as tighter risk management or regulatory constraints).
On the supply side, one apparent trend is that market-makers are focusing on activities that require less capital and less willingness to take risk. In line with this trend, in many jurisdictions banks say they are allocating less capital to market-making activities and are trimming their inventories, particularly by cutting holdings of less liquid assets.
In addition, proprietary trading (ie position-taking for purposes other than market-making) has reportedly diminished or assumed a more marginal importance for banks in most jurisdictions. Expectations are for banks’ proprietary trading to generally decline further or to be shifted to less regulated entities in response to regulatory reforms targeting these activities (Duffie (2012)). Overall, though, such a wind-down of proprietary trading will tend to limit market-makers’ ability to redistribute risky positions. Combined with reduced risk-taking in the financial system as a whole, this would then further reduce market-makers’ willingness to build up large inventories of less liquid assets
Who Makes Markets? Do Dealers Provide or Take Liquidity? This is a 2003 paper that looked at market liquidity issues pre crisis:
“Our main finding is that the contemporaneous correlation between weekly dealer order flow and stock returns is strongly positive. This implies that dealers do not provide liquidity on a weekly frequency….It also highlights the magnitude of the costs of allowing dealers institutional trading advantages. All the results strongly suggest that dealers are informed traders, since only informed traders should have positive price effects and such high excess returns driven by information. The institutional setup of providing dealers with low transaction costs, high transaction speed, and access to order flow information, is clearly valuable. As anecdotal evidence, one needs only to note the premiums paid for seats on the NYSE, the profitability of National Association of Securities Dealers Automatic Quotation (NASDAQ) Dealers, and the expanding role of ECNs.
These advantages create an indirect, but real, transfer of wealth from other market participants. If dealers are to have such valuable advantages, then they should be held to the socially beneficial function of providing liquidity, perhaps via narrow bid-ask spreads at reasonable depths at all times.
Automated Liquidity Provision – http://www.sec.gov/dera/staff-papers/working-papers/dera-wp-automated-liquidity-provision.pdf
Over the past decade, the task of liquidity provision has largely shifted from traditional market makers to proprietary automated systems that trade at high-frequency and across different exchanges and securities.
..the literature consistently reports that when liquidity providers are enabled so that they can quickly and frequently update their quotes in an automated way, prices are more efficient, liquidity is enhanced, spreads decrease, and adverse selection decreases – all of which are predicted by our model.