Liquidity means transactions cost; a highly liquid market is one where large transactions can be immediately done without incurring much transactions cost.
Where does liquidity come from? Many people look to "market makers" to create liquidity. The evocative term "market maker" seemingly suggests that they create liquidity. However, economic research has repeatedly demonstrated that liquidity comes from the economy and not from the market maker. It is people in the economy who wish to buy and sell that make liquidity for others. The market maker is a mere intermediary who facilitates the market, but he cannot create liquidity out of thin air.
The intermediation services of a market maker are often performed well by computers. The major equity markets of India are open electronic limit order book (OELOB) markets. This is the most modern form of market organisation in the world. This style of market dispenses with market makers, and allows people in the economy to directly trade against each other. Economic research sees a role for market makers when dealing with small-cap, illiquid stocks, but computerised order matching is an ideal way to deal with markets of higher volumes.
The OELOB market is built of limit orders. People place limit orders which are an intention to buy or sell, at a stated quantity and at a stated price. A limit order is a person who is saying "I'm willing to buy 100 shares at Rs.45 each". The limit orders are matched to generate trades, if mutually compatible, by the computer. E.g., when a computer finds another person saying "I'm willing to buy 100 shares at Rs.45 each", the interests of both buyer and seller are matched. The computer, tirelessly matching off mutually compatible orders, is a replacement for the market maker.
If a limit order cannot be matched immediately, it sits in the limit order book waiting for someone to hit on it. Hence, at any instant, one can go to the market and look at the entire limit order book, and know the trading intentions of all others on the market. The person who places a limit order obviously has the right (but not the obligation) to cancel the order.
A person who wants to be a market maker could easily give out "two way quotes" by putting a limit order to buy coupled with a limit order to sell. In this fashion, the limit order book market subsumes the human interactions of markets with market makers.
People can also place market orders which buy or sell "at the best available price". Market orders are matched against the best available limit orders at that point in time.
When a user approaches the market with an intention to trade, the market possesses liquidity solely by virtue of limit orders which exist, available for him to use them. If he wishes to buy, he would hit on existing limit orders which wish to sell, and vice versa. Table 1 is an an example of an order book for one stock at some point in time:
Id. Quantity Price 1 1000 3.30 Buy 2 1000 3.30 Buy 3 2000 3.40 Buy 4 1000 3.50 Buy 5 2000 4.00 Sell 6 1000 4.05 Sell 7 500 4.20 Sell 8 100 4.25 Sell
This shows eight limit orders waiting on the screen. An important feature of the OELOB market is anonymity: i.e., all that is seen is the prices and quantities that other traders are willing to trade at. The identities of the people placing those orders is not public. There is no room for coalition-forming, strategic games, and the cartels that afflict non-anonymous markets.
In this order book, there are four buy and four sell orders (though, obviously, their numbers do not always have to be equal). The distance between the best buy and the best sell (in this case, Rs.0.5) is called the bid-ask spread. If a person went into the market and tried to buy 100 shares, he would place a market order for 100 shares, this would be matched by the computer against order #5, the best available sell, and he would get the shares for Rs.4 each. If he tried to place a market order for selling 100 shares, he would be matched by the computer against order #4, and the shares would be sold at Rs.3.5.
Hence a person buys 100 shares, and turns around to the market and immediately sell them off, he is poorer by the bid-ask spread. The spread is the transactions cost which the market charges anyone for the privilege of trading (at a transaction size of 100 shares). High liquidity at 100-share transactions is synonymous with tight spreads.
Suppose a person wants to buy or sell 3000 shares. The market order on the sell side will hit on orders #3 and #4, and the market buy order will hit on #5 and #6. The "bid-ask spread" is misleading insofar as it only conveys the prices faced in doing small trades; for larger trades, the prices faced would be quite different.
This brings us to the concept of impact cost. We start by defining the ideal price as (3.5+4)/2, i.e. 3.75. In an ideal infinitely-liquid market, we would be able to do very large transactions on both buy and sell at prices which are very close to Rs.3.75. In reality, the market requires us to pay more than Rs.3.75 when buying and pays us less than Rs.3.75 when selling. The percentage degradation that is forced on us is called impact cost. Impact cost varies with transaction size. For example, in the above order book, a transaction of selling 4000 shares goes through at Rs.3.4. This is 10.3% worse than the ideal price of Rs.3.75. Hence we say "The impact cost faced in buying 4000 shares is 10.3%".
Impact cost is the most practical and realistic measure of market liquidity; it is closer to the true cost of execution faced by a user as compared with the bid-ask spread. This entire discussion, of course, highlights the fact that data for trading volumes is quite unilluminating as far as measuring liquidity goes.
In a speculative transaction, a person buys, holds for some time, and then sells. Each leg of this transaction suffers an impact cost, so the speculator loses twice the impact cost (if the trade size was only 100 shares, the speculator loses the spread). Hence, speculative volumes are attracted to markets with low impact cost. As compared with the other costs involved in doing a speculative transaction (e.g. the 0.009% that NSE charges each leg of each transaction), the spread (which averages 0.15% for the most liquid stocks of the country, and could easily get much bigger) is enormously bigger.
We should emphasise (a) that impact cost can vary between buy and sell, and can vary for every different transaction size and (b) impact cost fluctuates from moment to moment as traders change the limit orders that they have kept in the limit order book. However, at any instant in time, the stock of orders waiting in the limit order book embodies the liquidity that exists on the market.
When the entire limit order book is observed, impact cost can be calculated at any desired transaction size. A few examples of the impact cost seen on NSE are shown in Table 2 (they are averages for the impact cost seen in March 1997).
Security Rs.100,000 Rs.1000,000 Rs.2000,000 MTNL 0.143 0.271 0.389 TELCO 0.125 0.245 0.338 ACC 0.122 0.218 0.289 TISCO 0.099 0.162 0.209 Reliance 0.080 0.104 0.123 ITC 0.078 0.102 0.118 SBI 0.075 0.092 0.106
For example, these numbers mean that in March 1997, a market order for Rs.2 million of SBI faces an impact cost of 0.106% on average, but the same market order for TELCO faced an impact cost of 0.338% on average. Obviously, larger orders face a larger impact cost.
A market where the entire order book is displayed is highly transparent about liquidity. When all limit orders are publicly visible, people can measure impact cost of a transaction even before the transaction takes place, and use this to help them decide whether to do the trade. If the limit order book is not publicly visible, then trading is guesswork since the actual execution is always a surprise.
Transparency of liquidity cannot always be taken for granted. For example, the New York Stock Exchange (NYSE) is similar to a sluggish form of a limit order book market, but it is not open. The limit order book is the monopoly of one person called the specialist: users of the market are not allowed to see it. The transparency of liquidity on markets like the NSE and the BSE makes for efficient (computerised) analytical procedures in trading; when liquidity is less transparent, trading is perforce required to become more labour intensive and less scientific.
This approach also tells us that market liquidity comes about when the limit order book is thick and heavily populated by limit orders. People placing limit orders impart liquidity to the market. A market profuse with limit orders will have low impact cost. Building a liquid market is synonymous with attracting limit orders.
The discussion in this entire article has been couched in the language of trading shares. However, the ideas here, and the potential applicability of the OELOB market, extends into all financial markets. The dollar-rupee market, the treasury bills market, commodity markets, etc. can all fruitfully transition into a structure with anonymity, limit orders, computers matching orders, and transparency of liquidity.