List of Figures
List of Abbreviations
2 What drives Inventory?
2.1. Capital Constraints
3 How does Inventory drive Liquidity?
3.1. Inventory and Liquidity
3.2. Impact on the Bid-Ask Spread
4 How does Inventory affect Asset Prices?
4.2. Empirical Results
5 Changes in Market Structure
5.1. Dilution in the Role of Market Makers/Dealers
5.2. Technological Advances and High Frequency Traders
List of Figures
Abbildung in dieser Leseprobe nicht enthalten
Literature on market microstructure and the role of market makers has enormously increased in the last three decades. Market microstructure is concerned with the process of trading in the market in which market makers play a key role. Market makers are exchange specialists in continuous auction markets where securities are listed as well as dealers in the over-the-counter (OTC) market where securities are unlisted. Their role is to provide liquidity by standing ready to transact whenever buy and sell orders fail to arrive in the market at the same time. Specialists are designated market makers of the exchange who commit themselves to stabilize prices and to maintain liquidity and continuity in trading for generally five to ten securities in the market. To provide immediacy market makers set bid and ask quotes by observing the frequency of incoming buy and sell orders. Literature on market microstructure distinguishes between two types of models to explain how market makers determine bid and ask prices: inventory control models and asymmetric information models. Considering the latter, Bagehot (1971) explains there are three kinds of traders a market maker can be confronted with: firstly, traders who possess special information; secondly, “liquidity-motivated” traders who have no special information but want to liquidate securities into cash or convert cash into securities; and finally traders who act on information which has already been included in the price but which they believe has not. Market makers lose to the first kind of traders who are better informed and gain from traders of the second and third category. In order to stay in the market and generate profit, a market maker tries to set the spread between bid and ask price as to minimize losses and maximize gains from trading with different kinds of traders. Inventory control models based on Stoll (1978) and Amihud and Mendelson (1980) see bid-ask spreads as a compensation for market makers to provide inventory and to take on price risk by accumulating inventory. The market maker sets bid and ask quotes as to minimize risk and to maximize his utility. The effect of market maker inventories on prices can especially be observed in times of market stress, e.g. during the Flash Crash 2010 when an automated execution program sold 75,000 E-Mini S&P 500 futures contracts at a single blow. As a consequence of this large order execution, market makers accumulated a large net long inventory and tried to offset their position in the market which led to a 3% decline in price of the E-Mini in only four minutes. Due to the importance of inventories and the fact that asymmetric information models are extensively discussed in literature, this thesis exclusively focuses on inventory control models and provides a survey of theory and empirical results on the role of inventory in the price formation process. Because most of the relevant literature is based on the U.S. exchange market, this thesis is mainly confined on inventory control of specialists on the New York Stock Exchange (NYSE) and of dealers on the National Association of Securities Dealers (NASDAQ). It proceeds as follows. To understand the costs of holding inventory, Section 2 introduced three important drivers of inventory: capital constraints, liquidity and volatility. Section 3 summarises the effect of market maker inventory and its costs on liquidity and how this affects the bid-ask spread. In Section 4, the impact of inventory on asset prices, especially of inventory levels, is discussed in more detail. Section 5 briefly turns to changes in market structure and how they affect the role of traditional market makers and their inventories. Section 6 finally concludes.
2 What drives Inventory?
2.1. Capital Constraints
Amihud and Mendelson (1982) introduce a theory of inventory control in dealership markets. They argue that market makers are constraint by upper and lower inventory limits. These limits reflect a market maker’s risk aversion or his capital constraints imposed by institutions or by himself. Consequently, when a market maker faces capital constraints he is limited in his ability and willingness to accumulate absolute inventory and the costs of inventory are increasing with the absolute level of inventory. To avoid these higher costs, inventory adjustments are necessary. Shen and Starr (2002) add that with higher inventory, market makers’ financing costs are increased by financiers because of higher risk of insolvency. Tinic (1972) suggests that if market makers face capital constraints, they are not able to hold inventory for a long time period if they want to remain their purchasing capacity. Therefore, they will try to offset their positions at the end of each day. Brunnermeier and Pedersen (2009) show that financially constraint market makers are not willing to execute capital intensive trades where margins are high. Margins are set by financiers who lend market makers money on collateral to finance their trades. In each period margins can be changed, which imposes funding liquidity risk on market makers. When margins increase, their available capital is reduced, primarily for high-margin securities, and they are forced to reduce inventory. Gromb and Vayanos (2009) arrive at the same results as Brunnermeier and Pedersen. Huang and Wang (2010) further argue that high margins and capital requirements rise “participation costs” for market makers. Costs for rising capital and for adjusting positions or risk control and capital regulations imposed on the market maker reduce his inventory by constraining his capital and make participation in the market expensive or even impossible. This can lead to mismatches or imbalances in trades and in severe cases to dealer exits from the market.
The empirical study of Madhavan and Sofianos (1998), based on NYSE data, provides evidence that when market makers face inventory limits due to capital constraints they will adjust their positions to avoid hitting these limits. It is also observed that constraint market makers will participate more in smaller trades and larger capitalisation as well as trade less if trade size increases. Empirical results from Hendershott, Moulton and Seasholes (2007) confirm that with less “risk-bearing” capital of NYSE specialists, there is less willingness to take on inventory, especially for high-volatility stocks. Risk-bearing capital is the difference of total capital less the capital invested in inventory. Consequently, when risk-bearing capital is low, inventory and risk is supposed to be high. Comerton-Forde et al. (2010) follow the result that balance sheets of market makers can explain their willingness of providing liquidity by accumulating inventory. If a market maker owns large capital the effect of capital constraints on their inventory will only be small. Therefore, institutions can generally execute larger orders than individual market makers without facing higher inventory financing costs.
As mentioned above, the role of a market maker is to provide liquidity by acting as counterparty for incoming orders which cannot be matched directly. Therefore, market makers have to accumulate inventory, either long or short. However, Smidt (1971) finds that NYSE specialists are more willing to accumulate inventory when liquidity is already high. According to Tinic (1972) this can be explained by market makers costs of holding inventory and by their need to offset positions as soon as possible to avoid these costs. In liquid markets the probability of rebalancing inventory is greater and therefore it is more profitable for dealers to accumulate actively-traded stocks. However, these stocks usually do not need liquidity services by market makers. Instead, they have to manage large positons of thinly traded stocks, which increases their holding period of inventory and consequently their holding costs. Benston and Hagerman (1974) agree that costs of inventory decrease with the number of transactions a marker maker is able to execute due to the higher probability to lower his position level. The model of Ho and Stoll (1983) shows that if dealers face higher risk on offsetting their positions because of illiquidity in public markets they will engage in interdealer trading where they can rebalance inventory among each other. In the context of capital constraints, Brunnermeier and Pedersen (2009) as well as Gromb and Vayanos (2009) argue that because of higher risk in illiquid markets margins and financing cost of inventory increases and market makers reduce their positions.
The empirical results of Hendershott, Moulton and Seasholes (2007) confirm that illiquidity predicts less “risk-bearing” capital and therefore less accumulation of inventory. Panayides (2007) prove evidence that when there is need for liquidity, NYSE specialists accumulate inventory without seriously trying to rebalance their positions to avoid losses. When markets are liquid and their immediacy service is not needed they will rebalance their inventory level and try to offset losses i.e. generate profit. Cohen et al. (1977) support this result by observing that revenues from NYSE and American Stock Exchange (AMEX) market makers are smaller for thinly traded securities than for highly traded issues. Often they even appeared to be negative. Amihud and Mendelson (1991) and Grossman and Miller (1988) verify the effect of liquidity by examining the stock market crash in October 1987, better known as “Black Monday”. On that day, market makers were not able to absorb the enormous imbalances of order flow and the need for liquidity. Consequently, they were neither able to offset their huge positions nor willing to take on more inventory which led to a closing of markets. Some market makers completely exited the market because they were not able to finance further immediacy service.
Stoll (1978) explains that price volatility affects inventory through its holding costs. The model shows that holding costs are higher in volatile markets because while the dealer is waiting to offset his positions the probability of a decline in inventory value is higher. Therefore, he might have to sell stocks at lower prices and loses money. Furthermore, he argues that inventory holding costs of dealers do not only depend on the volatility of one stock but also on the correlation of a stock with the total inventory. Bradfield (1979) argues that if a designated market maker acts according to his obligation, he uses his inventory to buy stocks when prices are low and to sell when prices are high. Therefore, he lets his inventory fluctuate and accumulate positions when there is price volatility. However, because of overnight costs of carrying inventory from one trading day into another, the market maker tries to end the day at a desired inventory level. Adjusting inventory according to prices as required by the exchange generally does not lead to the preferred level. Consequently, price volatility limit inventory management and can lead to higher costs for specialists. O'Hara and Oldfield (1986) and Shen and Starr (2002) arrive at the same conclusion. In the context of capital constraints Brunnermeier and Pedersen (2009) argue that higher volatility increases margins i.e. financing cost and therefore limit inventory.
Madhavan and Sofianos (1998) prove that despite higher holding costs there is a positive relationship between price volatility and participation of NYSE specialist in trading. Tinic and West (1972) as well as Benston and Hagerman (1978) found for OTC markets that if a dealer has a diversified portfolio, price volatility has less impact on costs. However, diversification is expensive itself and also restricted by the exchange.
It has to be taken into account that the effect of capital constraints, liquidity and volatility on inventory cannot completely be separated. Pagano (1989) suggests that illiquid stocks are also generally more volatile than liquid issues. Furthermore, Brunnermeier and Pedersen (2009) show that in high volatile markets or illiquid markets also cost of financing inventory are higher and therefore less inventory will be accumulated.
3 How does Inventory drive Liquidity?
3.1. Inventory and Liquidity
As already mentioned, market makers provide liquidity by accumulating inventory, either long or short, to absorb demand of immediacy. According to Grossman and Miller (1988) this adjustment of demand and supply over time leads to an equilibrium level of liquidity. Literature generally agrees that if a market maker is not willing or able to accumulate inventory, liquidity will decline and if a dealer is able and willing to accumulate a high level of inventory, liquidity will increase. Their willingness and ability to provide liquidity depends on holding costs of inventory and its drivers. Three important drivers: capital constraints, liquidity and volatility were already discussed in section 2. Demsetz (1968) introduces a transaction cost theory which states that the main factors of liquidity cost are the time rate of transactions, the price, and the number of stock exchanges on which the issue is traded. Tinic (1972) completes this theory by arguing that holding costs of inventory can be divided into three main categories: one; the costs of positioning in an issue under certainty and uncertainty; two, the cost structure of the market maker unit and three, the costs of profit margin changes. If there is no uncertainty, holding cost of the first category consists of the stock price, the number of issues and the length of the time period shares are carried, which influence borrowing costs of capital. High borrowing costs result in lower inventory and therefore lower ability to provide liquidity (see also section 2.1). Considering uncertainty, the volatility of a stock also has to be taken into account (see section 2.3). Because holding cost in high volatile markets are high, dealers are less willing to provide liquidity. Furthermore, the level of trading activity of the market maker, which depends on the liquidity in the market, influence the cost of inventory (see section 2.2). In illiquid markets, the probability of being able to efficiently balance inventory is lower. Therefore, illiquid markets tend to be more costly for market makers who limit their provision of liquidity. Finally, the number of institutions holding the stock and their relative holding size of the issue also affect inventory cost. If the institutional concentration is high, the probability that institutions can offset their large positions among each other is small. Therefore, a market maker needs to take on larger positions and faces higher costs. This can also lead to lower participation of dealers in those markets. The second category consists of the purchasing capacity of a market maker and of the size of his portfolio. His purchasing capacity depends on his capital and on the ability to borrowing capital (see section 2.1). If market makers are not capital constraint they are able to accumulate more inventory and consequently provide more liquidity. The number and size of stocks a dealer manages is also important for liquidity. If he manages more than one stock, he can diversify costs. However, through less specialization in each stock and larger positions, his flexibility and efficiency to offset inventory can diminish. This increases risk and lowers the quality of providing liquidity. Following Tinic’s theory, this would imply that specialists have lower cost of inventory and provide more liquidity than dealers. The last category of holding costs refers to constraint price setting power of market makers, which lowers their profit. Such constraints can be exchange regulations and intermarket competition. Low profit margins means higher costs for the market maker and therefore less willingness to accumulate inventory and provide liquidity. Lo, Mamaysky and Wang (2004) join Tinic (1972) by saying that higher costs of holding inventory will lead to less frequently trading of market makers or to no participation at all, which has a negative effect on liquidity. Huang and Wang (2010) also argue that if there are higher liquidity or participation costs for a market maker, like cost of rising capital or risk control, he will provide less liquidity or even exit the market.
Gromb and Vayanos (2009) and Brunnermeier and Pedersen (2009) agree that liquidity increases with the size of the market maker’s capital and inventory. With less capital available to finance inventory, liquidity declines in all markets where the market maker is involved. This is especially true for volatile and high-margin markets. They argue that because higher volatility increases margins and limits inventory, market makers are less willing to participate in volatile and illiquid securities. This sets up a “flight to quality” i.e. a flight into larger and less risky securities, and even decreases liquidity in already volatile and illiquid securities. In times of very high market volatility, a “flight to quality” can even lead to a liquidity crisis. Grossman and Miller (1988) similarly argues that the larger the group of market makers, and therefore the capacity to absorb demand of immediacy, the higher market liquidity.
The model of Stoll (1978) and Ho and Stoll (1981) suggests that cost of liquidity depend on stock volatility, risk aversion of the market maker and transaction size. However, Amihud and Mendelson (1982) argue that price risk can be diversified by the specialist’s organization and is therefore no reason for higher inventory costs and lower liquidity. According to their model, liquidity is affected by inventory financing costs and capital constraints.
 U.S. Securities and Exchange Commission, 2012, Specialists, http://www.sec.gov/answers/specialist.htm.
 Cf. Bagehot, 1971, pp. 13-14.
 Cf. U.S. Commodity Futures Trading Commission, U.S. Securities and Exchange Commission, 2010, pp. 2-3.
 See also Demsetz (1986).
 See also Vayanos, Wang (2012).
 See also Hendershott, Moulton, Seasholes (2007) and Chordia, Roll, Subrahmanyam (2000).
- Quote paper
- Evelyn Rill (Author), 2015, The Role of Market-Maker/Dealer Inventories in the Price Formation Process, Munich, GRIN Verlag, https://www.grin.com/document/309403