Adverse selection is the phenomenon that says when entities such as banks are faced with a price to insure against default, those that find the price cheap are most likely to need the insurance and buy it, while those whose loan book is less risky won’t buy the insurance.
The provider of insurance (or for that matter a hedge against an adverse price movement or even trying to price the original level of risk), will look at the expected probability of default across the risky entities and other things being equal, price the value of the risk as an average between the two groups. As a result, they end up insuring (or lending to) a higher proportion of those entities that have a higher risk of default than originally intended – the people that need the insurance the most are the greater proportion of customers.
There are many pricing methods to get high- and low-risk entities to self select -- schemes that separate the good from the bad. Most of these, in the face of competition from buyers and sellers of risk, and insurance, break down. In the complex world of derivates, hedges and CDS, such price schemes can be undone and are surprisingly vulnerable to adverse selection.
Should the default event occur, adverse selection leads to a higher than expected payout on insurance claims, and the insurer itself is at risk of insolvency, as we discovered with AIG and the domino effect of the sub prime mortgage crisis. Today, markets are highly concerned about the same effect of the current euro crisis.
In the face of adverse selection, raising the price of insurance (or loan) does not solve the problem. It simply divides the proportion that “under insure” (because the price is too high relative to their own risk profile) and those that over-insure (because it is cheap relative to their own risk profile) to higher levels of default probabilities.
Rather, the higher the price of insurance the more likely it is that entities needing it the most form an increasing proportion of the insured. As a result, when the price of risk and cost to de-risk rises, it tends to do so exponentially; insurers increasingly fear their exposure to adverse selection. Prices spike.
One mitigation strategy is for the insurer to not offer the full cover against the risk of default – the counterparty must still bear a proportion. Hence, the underlying contract (bond, loan) to hold that risk, and the price of insuring against default, rise in parallel and exponentially when events move closer to the likelihood of default.
So what is the relevance of adverse selection today?
The phenomenon helps to explain price spikes and volatility in times of uncertainty. We have all at some stage seen the graphs of most Euro zone sovereigns’ yields or CDS spreads over time, with exponential blow-outs following a material but not excessively dramatic change in risk.
People can relate to the fear associated with escalating prices for holding risk assets and their corresponding price of insurance. Intuitively, they ask, "What if the return or insurance cover is not enough to reflect the underlying risk?” Indeed, with adverse selection, it may never be.
In such markets, price discovery that reveals the extent of adverse selection related values, only occurs in two circumstances: If the credit event (default) occurs, then people find out whether they had enough insurance or not and insurers find out whether the risk of default matched the risk they insured. Both are ex poste price discoveries so no use in evaluating whether the premium or risk reward includes any distortions.
The second point of price discovery is when credit and insurance markets “freeze” at the peak of a price spike. At that point, yields and corresponding insurance premiums are so high that even those with the highest exposure refuse to buy because they believe the prices are too high – bond values hit a floor and begin to recover, without the default event occurring.
In this view of the market, adverse selection and fear of it drives price spikes, but these form an important but unpleasant price discovery mechanism to either force a default or hit a top where all buyers and sellers no longer participate in the market, before beginning to revert. The vicious cycle turns virtuous; the bears get overtaken by the bulls.
Such a price discovery method may increase the risk of default, for two reasons. Firstly, as the price of risk and premiums to protect against bad outcomes blows out, it adds to the cost of staying solvent for all parties involved. This additional cost for a given level of funds available to service the debts and buy insurance, may create a liquidity shortage which in turn causes a cash solvency crisis.
This is the fundamental concern with the European sovereigns themselves. The higher the cost of debt the less likely future economic performance will be sufficient to meet the debt and so the cost of borrowing rises even more; a simple vicious cycle that stops when adverse selection gets priced out of the market, or a default actually occurs.
Secondly, as prices spike exponentially, a 1% margin of error in estimating how much insurance is required could lead to much higher than 1% under- or over-insurance, with the risk being more to the downside (i.e. under insurance). In other words and ironically, the companies that believe they are covered when they are not are those whose margins of error may lead them to a solvency crisis, not the entities that thought they were over-exposed to the risk and kept buying insurance or hedged that risk even as premiums rose.
Despite the potentially painful imperfections of price spikes, vicious cycles and frozen liquidity markets, they have their value as mechanisms for discovering the upper limits to prices for risk in the face of adverse selection.
The current eurozone crisis has a number of additional complicating factors that have undermined this market function. Firstly, in an attempt to preserve stability (prevent a vicious cycle), central banks are intervening with the unintended consequence of preventing the only value of the vicious cycle – the price discovery part.
Perhaps it is because they are aware that bank defaults are about to occur and don’t want the horrendous fallout from such an event. Or perhaps it is a signal that if the price of risk and the cost to de-risk went higher, it would lead to a definite default? Or perhaps it is because a bank did default, and would have, if not for the intervention?
The second problem is the implied recasting of the role of central banks from lenders of last resort to holder of first default. Italian and Spanish bond buying by the ECB for example, has four probable endings:
- They sell the bonds at profit in the future when yields “normalise” (money supply would decrease as a result)
- Hold till maturity and recover the face value, earning the higher yields along the way.
- Write down the value of the debt in the event of default, ultimately by printing more money to pay for it since by definition central banks cannot default themselves.
- Pass the risk onto a third party (EFSF, ESM, future Eurobond issuer), sterilizing the inflationary impact of the initial bond buying project.
Intervention itself creates a moral hazard – a potential expectation in markets for further intervention as soon as things go off course, and a revaluation of the risk of holding assets.
But it also creates the problem of preventing risk and insurance markets from freezing as liquidity dries up at prices quoted. Because in doing so it eliminates the only way markets know how to evaluate the possible extent of adverse selection using market prices as the discovery mechanism.
Imagine running an eBay auction where you are selling a used computer, and the person can buy insurance against breakdown. As the auction nears its end, the prices tend to spike. Liquidity dries up as those believing the probability of the appliance working and being fit for purpose are the ones making the highest bids.
Now imagine that eBay steps in and says “there must be at least ten bidders in the market to set a fair price.” Price discovery halted, the pricing mechanism distorted and a whole new set of conjectures arise about what any given price actually represents and what moral hazard/ new incentives it creates.
That sort of ex ante intervention is very different (and price-distorting) compared to ex-poste intervention by eBay, when, for example, the buyer never receives the computer after successfully bidding and paying for it.
Yet, central banks justify the ex ante intervention on concerns for stability of the monetary policy transmission mechanism (the banking system). My argument here is that, once in a while, as painful as it is, that system has to freeze to work out the true price of risk. By intervening in anticipation of that freeze, rather than ex-poste, central banks inadvertently destroy the price discovery value of the freeze and prolong uncertainty rather than eliminate it.
Following the coordinated offer of foreign currency swaps by the major central banks, are you more or less sure that a default is likely to occur? Following the buying of Italian, Spanish and Greek debt by the ECB, are you more or less sure that those countries are at risk of default?
Implications For Your Investment Strategy
With price discovery temporarily suspended by the central banks and now a significant chance of intervention occurring again at some random price point in the future, one set of uncertainties and has been replaced by another. Yet the adverse selection problem has not gone away. Price spikes have not gone away.
How far down or up yields, CDS spreads, commodity and stock prices rise and fall will have to go before markets freeze out the adverse selection problem, remains to be seen. How many defaults along the way remains to be seen.
In other words, price uncertainty, extreme price volatility and lack of transparency in what the price of an asset represents is now perhaps made more complicated with the additional variable to consider: When could the next central bank intervention be and what will that imply about its changing role? Or perhaps the latest intervention was not a change of role: sSme entity on the brink of default made the call to a central bank, prompting their rapid intervention to prevent a 2008-style domino effect.
Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.