Market Liquidity: A Primer - Brookings Institution

Market Liquidity: A PrimerBrookings 2015conditions. At other times, volatility is affectedby changes in bid/ask spreads or otherelements of liquidity. When it is more expensiveor harder to trade, then fewer traders arewilling or able to step in when prices move outof line by modest amounts, allowing prices toswing more widely.Direct transaction costs for investors rise. Insome cases, external factors, such as increasesin regulatory requirements for trading, directlypush bid/ask spreads higher, which raisestransactions costs for investors, which is oneaspect of liquidity. Further indirect effects resultfrom cutting transaction volumes, which mayalso lengthen the time necessary to complete atransaction.Whatever the cause of increased volatility, itgenerally reduces the return for investors whoare buying or selling in any significant size, astheir initial purchases or sales will move themarket price further in the wrong direction forthem.In other cases, the causality runs in the otherdirection, and markets initially become lessliquid in some other way, such as through a risein the volatility of price movements. Bid/askspreads would then usually increase as well, forseveral reasons. Transaction volumes wouldtend to fall, so the dealer’s fixed costs would bespread over fewer transactions, raising the costper transaction. Further, risk premiums wouldrise as well to cover the higher price volatility,as may also be true of the capital and liquiditycharges, at least if illiquidity persists.There is greater potential for financial crises.Illiquidity in financial markets can help trigger orexacerbate a financial crisis by creating actualor paper losses at banks or other financialinstitutions. If a bank needs to raise cashquickly, perhaps to meet deposit outflows inthe event of a loss of confidence in thatinstitution, they will likely need to sellsecurities, especially if they have an excessivemismatch between the maturities of theirassets and liabilities. In illiquid markets, thiswould require “fire sales” in which the selleraccepts a significantly lower price in order toget cash quickly. In addition to the direct loss tothe troubled institution, which may threaten itssolvency, rapid declines in securities prices canaffect other institutions, either because theytoo need to sell or because they use “mark tomarket” accounting for their assets andtherefore paper losses directly affect theircapital positions.Whatever the derivation of the highertransaction costs, they flow through to lowerreturns for investors when they buy or sell theinstrument.Volatility of prices increases. The biggestfactors moving securities prices are those thataffect perceptions of their fundamental value,such as good or bad news about a firm’screditworthiness or an overall move in interestrates. However, the rapidity and extent of pricemovements is also influenced by marketliquidity. If there are many potential buyers andsellers and they can transact quickly, easily, andcheaply, then price movements tend to besmoother as news events are factored intoprices quickly based on the market consensusabout their significance. Similarly, if a marketparticipant wants to buy or sell a large block ofbonds, they can do so without greatly movingthe price.Bond prices fall as Investors demand higherliquidity risk premiums. When investors decidethe minimum interest rate they will accept on abond, they take account of multiple factors.First, they need a base return that compensatesthem for giving up the use of their funds untilthe maturity of the bond, often known as the“time value of money.” Second, they need to becompensated for credit risk, the possibility thatthey will not be repaid in full. Third, they maycharge an interest rate risk premium to reflectAs with transaction costs, sometimes volatilitydirectly changes, perhaps due to higheruncertainty about economic or monetary policy4

Market Liquidity: A PrimerBrookings 2015the potential for a decline in value if interestrates rise. Fourth, they will charge a “liquiditypremium” based on the degree of difficulty orcost they will encounter if they decide to selltheir investment early. (On top of these basicelements, there may be others, such as foreignexchange risk premiums, depending oncircumstances.)global financial crisis of 2007-9. However, thereis a great deal of controversy about the extentto which this has occurred and whether itrepresents a bad thing or a return to normalconditions after an unsustainably high degree ofliquidity in markets. (Fender at. al., 2015,provides a good overview of the changes sincethe crisis.)If markets become less liquid, then investorsover time should increase the liquidity riskpremium that they demand, raising their overallrequired interest rate. This would cause theprice of existing bonds to fall, since lower pricesare needed to raise the effective interest rateon the amount invested.The picture looks different depending on whichaspect of liquidity one focuses on and whichmarkets one considers.Level of dealer inventories. This is not a directmeasure of liquidity, but rather an indicator ofthe potential for dealers to provide liquidity.Large inventories make it easier for a dealer tosupply bonds if customers desire them. Theyalso tend to be an indicator of the willingnessand desire of dealers to make markets. With theexception of government bonds at that nationallevel, dealer inventories are down pretty muchacross the board in the last few years. Thedecline was very substantial in many types ofbonds, particularly corporate bonds. Sovereignbonds have shown little decline, but a large partof this is likely due to new regulatoryrequirements and other pressures to hold largevolumes of government bonds at the banks andmajor dealers. The chart below provides somedata on inventories from Fender, et. al., 2015.Capital raising becomes more expensive.Similarly, an increased liquidity risk premiummeans investors would demand higher interestrates when businesses and governments issuenew bonds. This would directly flow through asa cost to borrowers, including householdswhose borrowing is financed indirectly throughfinancial markets, such as is true for mostmortgages.Has market liquidity actuallydeclined?Almost everyone believes that market liquidityhas fallen overall since the period prior to the5

Market Liquidity: A PrimerBrookings 2015transactions in many market segments, whichmay indicate that investors have found theneed to break up their transactions due to theinability or high cost of moving large blocks in asingle transaction. However, there could beother factors leading to this reduction in size,including the rise of trading strategiesemploying frequent trades in smaller sizes to tryto profit from fleeting arbitrage opportunities.Bid/ask spreads. On the other hand, there hasbeen much less of a movement in bid/askspreads. On this basis, one would not presumethere were any concerns about a decline inmarket liquidity, except in certain marketsegments which were already less liquid. Thechart above, also from Fender, shows corporatebid/ask spreads over time.Volatility on “normal” days. Price movementshave been relatively calm for the most part,again not indicative of a current problem withliquidity.Time to completion of transactions. Althoughthere do not appear to be good measures,anecdotal evidence suggests some slowingdown of the disposition or acquisition of largepositions. This would be consistent with smalleraverage transaction sizes.Bouts of extreme volatility. On the other hand,there have been a few incidents, described inthe next section, in which price movementshave been extreme enough to trigger fears thatmarkets have indeed become less liquid.Overall, the decline in liquidity has been mostmarked in riskier market segments, asdemonstrated above in the charts from Fender,which showed little or no decline in the liquidityof government bond markets of developedAverage size of transactions. There hasgenerally been a decrease in the size of6

Market Liquidity: A PrimerBrookings 2015economies, but noticeable declines in liquidityof corporate bond markets. This results bothfrom factors specific to such markets (includingchanges in regulatory requirements that raisethe required level of capital for banks anddealers holding some instruments) and as areflection of an overall “flight to quality” bymany bond investors after the financial crisis, aswell as from various other factors.after the testimony. This may not seem large,but is quite a sharp move for a governmentbond market.October 15, 2014 Treasury market rally. Thisincident is so complex, and the causes sounclear, that it still does not have a singlenickname. A variety of factors led to a rise inTreasury market prices roughly equal to theentirety of the taper tantrum within about onehour, with prices subsequently gyrating stronglyover the remainder of the day. US authoritieswill soon conclude a study of this episode thatmay shed more light on the underlying causes.Similarly, the IMF’s Global Financial StabilityReview of October 2014 highlighted concernsabout market liquidity particularly in the highyield bond and emerging market bond areas.Swiss Franc revaluation. For several years, theSwiss central bank held down the value of theSwiss franc versus the euro, in order to mitigatea loss of competitive position by Swissexporters versus those in the eurozone. Thisrequired the Swiss to buy large sums of euros inexchange for francs. Eventually, the holdings ofeuros grew very large, as did the likelihood thatthe central bank would eventually have to takea loss on these holdings, in part due to theanticipated advent of Quantitative Easing by theEuropean Central Bank. As a result, in Januaryof 2015, the Swiss National Bank gave up andallowed the Swiss franc to rise, switching to apolicy of intervening sporadically if marketforces appeared to be excessive. This retreat bythe central bank caused the Swiss franc to rise30% in the first 13 minutes, with knock-oneffects in other foreign exchange markets. (Thefranc gave back some of these gains over thecourse of the day, but most of the initial impactremained.) Some observers believe that thespeed and extent of the initial price movementwould have been considerably less in moreliquid markets. This is hard to judge asdeveloped economies rarely undertake this kindof capping of foreign exchange rates anymoreand therefore it is difficult to compare withother instances where such a cap wasunexpectedly withdrawn.What do the recent bouts ofmarket volatility mean?There have been at least four occasions in thelast several years that may indicate a greatervulnerability of fixed income markets to periodsof excessive volatility as a result of regulatoryand other market changes 1. As noted below,however, each of the episodes were associatedwith major news events, often related tocentral bank activities, that make it difficult topin down what portion of the volatility was“excessive” and what was a reasonableresponse to fundamentals. These are:The “taper tantrum”. When then-ChairmanBernanke testified before Congress in May 2013that the Fed might “taper” off its purchases ofbonds in the markets more quickly than some inthe markets had expected, there was a quickmovement down in government bond prices,which carried over to most other categories ofbonds, which generally price on the basis of aninterest rate spread over the government bondrates. The 10-year Treasury bond saw its marketprice fall about 3% in the course of two days,with most of this occurring in the first few hours1Some analysts who have reviewed this paper in draftformat have nominated additional examples, such as sharpprice movements in Japan connected with theannouncement of changes in the investment strategy ofthe government pension plan.7

Market Liquidity: A PrimerBrookings 2015Volatility of European government bonds inearly 2015. Prices of government bonds in thecore of Europe fluctuated sharply in the firsthalf of 2015, with a cumulative move in German10-year government bond prices of 7-8% frompeak to trough. Within this overall trend, therewere fairly rapid moves on some days. Someascribe the sharpness of the moves tounderlying liquidity problems, although theargument is less strong than in the case of thetwo incidents involving US Treasury bonds.which can push liquidity in either directiondepending on the particular circumstances.The first compelling reason is that marketliquidity in the US was greater in the run-up tothe financial crisis than it had ever been, quitesubstantially so in many markets. In part, therewas clearly a self-reinforcing cycle of increasedliquidity leading to lower liquidity risk premiumsdemanded by dealers and investors, leading tostill more liquidity. A second major componentwas a belief in the “great moderation,” thatcentral banks had determined how tosubstantially reduce volatility in the economyand consequently in financial markets. Lowervolatility begets greater liquidity as dealers andinvestors become more willing to take positionswithout fear of excessive losses. Both of thesefactors have vanished or reversed, helping toexplain the lower liquidity levels today.Outside these markets, there was also the“flash crash” in the stock markets in May 2010and smaller versions since.At the end of the day, it is difficult to tell howmuch meaning to ascribe to these events. It iscertainly possible that they represent the tip ofthe iceberg and that once we return to morenormal economic and monetary conditions,these types of volatility events will be morefrequent and potentially much more painful.However, it is also dangerous to generalize toomuch from a few data points. One couldcertainly argue that at least some of the eventsmerely showed the market reacting sensiblyand swiftly to new economic news, such as thewithdrawal of the Swiss central bank as aprovider of massive artificial support to theeuro or the news about the Fed’s intentions forits future bond purchases.The second compelling reason is that thedealers who have dominated fixed incomemarket making are virtually all subject to awhole set of new regulations that make it moredifficult and more expensive to provide thatservice. It would be surprising if such a distinctdeterioration in their business position did notlead to a significant retrenchment and repricingof their liquidity provision to the markets. Asnoted earlier, dealer inventories in mostmarkets have come down quite markedly, inline with this expectation.It is probably best to view these incidents as redflags, and indicators of the degree to whichvolatility might become more normal, ratherthan drawing stronger conclusions from thislimited set of data points.There are quite a number of new regulationsthat have a significant impact on the cost ofdoing business as a market maker:Basel III capital accord. The Basel Committeeon Banking Supervision is the globalcoordinating body for bank regulators. Althoughit cannot directly bind national governments, itsrules are virtually always adopted, sometimeswith modifications. The Basel Committeepromulgated the third version of the BaselCapital Accords after the financial crisis andthey are well along the phase-in process today.The latest version significantly raises theWhy would we expect marketliquidity to be down?There are two broad and compelling reasons toexpect market liquidity to have declined,especially for securities that were already lessliquid. In addition, there are a number of otherfactors at work that are of lesser significance or8

Market Liquidity: A PrimerBrookings 2015intended to ensure that banks and majordealers have high levels of liquid assets to meetpotential demands for funds in a crisis and thattheir overall business models do not have anexcessively large mismatch between thematurity of their assets and their liabilities. (SeeElliott, 2014, for a primer on bank liquidityrequirements, which also apply to the majordealers.) The Liquidity Coverage Ratio (LCR) is astylized stress test to ensure that a bank has theability to handle a 30-day liquidity crisis in themarkets. Under this test, assets which arelonger-term or less liquid effectively need to befunded by longer-term liabilities, which tend tobe more expensive. This raises the cost ofholding inventories of most bonds.amount of capital required by banks and majorsecurities dealers, which makes it moreexpensive for them to do business. (See Elliott,2010 for a primer on bank capital.)Basel 2.5. The capital required for assets held ina bank’s “trading book” was considerably lowerunder Basel II than was the case for otherassets, such as securities that were intended tobe held to maturity. After the financial crisis,there was such a strong consensus that thesecapital levels needed to be raised sharply, thatnew rules were put in place to modify Basel II inthis area even before the Basel III accord wasagreed. (Hence, the nickname of Basel 2.5.)Trading book assets now require multiples ofthe capital previously mandated, representingone of the sharpest percentage changes incapital requirements. In addition, the BaselCommittee is currently conducting a review ofthese requirements and there is an expectationof still further increases.Net stable funding ratio. The second liquidityrelated requirement in Basel III is a ruleintended to ensure that banks and majordealers do not have an excessive mismatchbetween the maturity of their liabilities and thatof their funding. This produces a similar effectto the LCR, by raising the cost of funding forlonger-term instruments, such as most bonds.Leverage ratio. The Basel Committee alsoconcluded that its core approach, which usesrisk weightings so that more capital is requiredfor riskier assets and less for safer ones, was toosubject to gaming or error when used on astand-alone basis. Therefore, a “leverage ratio”has been adopted as well which, in essence,requires the same level of capital for all assets,regardless of risk. Banks must meet the higherof the capital levels required by the riskweighted approach and that calculated by theleverage ratio. In the US, regulators wentfurther and established a “SupplementaryLeverage Ratio” (SLR) for the largest banks thatis higher still. The SLR has particular impact ontrading, since most of the instruments that aretraded, or are used to hedge trading positions,involve securities with very low credit risk.These have correspondingly low capitalrequirements under the risk-based rules, but donot receive any benefit under the leverageratio.Single counterparty credit limits. The DoddFrank Act required that the rules be tightenedon the amount of credit exposure that thelargest banks and their affiliates could take toany one counterparty. Bonds in dealerinventories count against this limit as do manyof the instruments used by dealers to hedgetheir risk of holding those inventories. Thetighter requirements mean that the largestbanks have to ration their credit exposuresmore than they did, which adds an opportunitycost when dealing activity uses up some of thisroom under the exposure limits.The Volcker Rule. Banks and their affiliates arenow prohibited from engaging in “proprietarytrading”. As I, and others, have written aboutextensively, there is no clear meaning to theterm and therefore dealers have a strongincentive to cut back on some of their marketmaking that might be misinterpreted asproprietary trading. In particular, dealerLiquidity coverage ratio. Basel III also includestwo completely new requirements that are9

Market Liquidity: A PrimerBrookings 2015inventories that rise too much or too qui

securities dealers to act as market makers. (These rules include the liquidity coverage ratio, the net stable funding ratio, the sup