Fundamentals of Liquidity
Fischer Black
Graduate School of Business
University of Chicago
June, 1970
In a world in equilibrium, it simply does not make sense to think of
liquidity as an attribute of an asset that affects its expected return.
There is no trade-off between liquidity and return.
In the current world, there may be, for institutional reasons, such Q
trade-off. But it would be very difficult to establish laws governing
the trade-off, since attempts of individuals and institutions to get
around the institutional constraints will tend to change the terms of
the trade-off. And since there will be no trade-off in equilibrium, we
would need very powerful evidence to be convinced that the world is
consistently out of equilibrium in such a way that more liquid assets
have lower expected returns than less liquid assets.
LIQUIDITY AND REAL ASSETS
It is often claimed that some real assets are more liquid than others,
because there is an established market for such assets, and they can be
sold in this market on short notice.
A real asset has two functions: it acts as a store of wealth, and it is
an object that has value in use. In theory, these two functions can be
separated, as when one person owns an asset, and another person rents it
for use. In practice, it may be advantageous for th~ user to mm the asset,
-2-
since he then has an incentive for keeping it in good condition.
It is true that the selling costs are lower for some assets than for
others. However, selling costs should be assigned to the user of the
asset, not to the owner. Thus variability in selling costs should not
affect the return that the owner receives on the wealth he has invested
in the asset.
For example, suppose that the retail value of a new car is $3000.
Suppose that $300 of this is selling cost, so that its wholesale value
is $2700. After one year, suppose that its expected wholesale value
is $2000; and after two years, $1500. Suppose further that the uncertainty
in the value after one year or two is such that the required
return on wealth invested in a car is 10% per year.
Then the cost of using the car has three components: (1) the initial
selling cost of $300, which is independent of the time the car is used,
(2) the depreciation in the wholesale value of the car over the time
it is used, and (3) the required return on the amount of wealth that
could be made available at any time by selling the car at wholesale.
Thus the costs of using the car for one or two years are as follows.
Initial Required
Selling Return
Cost DeEreciation On Wealth
One Year: $300 $700 $270
Two Years: 300 1200 470
Total
$1270
1970
The important point here is that the owner of the car can rececve all
of the required return on wealth for assets of this degree of risk. He
does not bear the cost of transferring the asset from one user to another;
the user bears this cost. In a competitive world, wher? the person with
-3-
wealth can get the required return on his wealth by investing it in
financial assets, the extra costs of investing his wealth in a car
for his own use must all be assigned to the use of the car rather
than to the ownership of the car.
CASH FLOWS AND REAL ASSETS
The fact that liquidity should not affect the return on investment in
real assets becomes especially clear when we consider the relation
between the purchase or sale of an asset and the cash flow connected
with the purchase or sale.
There is virtually no relationship between the legal transaction in
which a buyer agrees to pay a certain amount to a seller, in return for
an asset owned by the seller or services to be provided by the seller,
and. the settlement of the transaction in the form of one or more cash
flows.
The seller can get most of the money that he will realize from the
sale before he actually makes the sale, by borrowing from a lender
in anticipation of the sale. On the other hand, the seller may provide
credit to the buyer, and may get his money in the form of periodic
_payments on an installment contract. Or a lender may give the seller
money before the sale, and collect from the buyer (in installments)
after the sale. Thus the seller can get his money either before or
after the sale, and the buyer can pay his money either before or after
the sale. Both the buyer and the seller can arrange to have the actual
cash flows occur at convenient times that bear almost no relationship
to the time of sale.
In particular, both the buyer and the seller can arr~nge for the cash
flows to occur at a future time that is fixed in advance. So both of
them can know well in advance when the cash flows will occur, and neither
-4-
has the problem of dealing with unexpected cash flows.
LIQUIDITY AND FINANCIAL ASSETS
Like a real asset, a financial asset may have more than one function.
In addition to serving as a store of wealth, a financial asset may make
it possible to transfer risk from one person to another, and may make
it possible for speculators to make "bets" on the fortunes of a particular
company.
But these functions are separable. There is no reason why the person
who supplies the money for a financial asset should take the risk
associated with the asset. And the risk can be transferred from one
person to another independently of any transfer of the money investment
from one person to another.
For example, government bond dealers hold large quantities of bonds
in inventory, financed almost entirely by bank loans and loans from
other financial institutions. The dealers are bearing the risk of
fluctuations in the value of their bond portfolios, but they are
supplying only 5% - 10% of the money that is invested in the bonds.
One dealer might sell a bond to another dealer, without any change
in the institution financing the bond; this would be a transfer of
risk without any substantial transfer of funds.
Consider a long term corporate bond. There is no theoretical reason
why a financial institution could not guarantee, for a fee, the payments
of interest and principal on the bond. He would expect the fee to be
related to the amount of market risk in the bond. Such a bond would
then be, from an investor's point of view, virtually free of risk of'
default. The transfer of the bond would be completely independent of
the transfer of the guarantee. A second institution might take over
the guarantee function from the first institution, without affecting
-5-
the bond holder. And the holder of the bond could sell it to another
holder without any change in the institution guaranteeing the bond.
It is clear that the risk of default on the bond, and changes in the
fortunes of the issuer, will affect the terms on which the guarantee
can be transferred. If there is a dealer in guarantees, these factors
may affect his spreads. But these factors will not affect the terms
on which the bond itself can be transferred. Nor will they affect
the spreads of bond dealers.
Similarly, there is no theoretical reason why a financial institution
should not guarantee a long term bond against fluctuations in price.
This would convert the long term bond into a short term bond, from
the holder's point of view, and would separate the function of supplying
the money from the function of bearing the risk of price fluctuations.
Again, it is clear that changes in interest rates will not affect the
terms on which the bond can be transferred, and will not affect the
spreads of a dealer in bonds. They will affect the terms on which
the guarantee can be transferred, however, and may affect the spreads
of a dealer in guarantees.
Instead of having a financial institution guarantee the bond against
price changes due to changes ~n interest rates, the corporation issuing
the bond might choose to have a coupon tied to current short term
interest rates rather than a fixed coupon. This would also serve
to eliminate the risk of changes in price due to changes in interest rates.
Thus a long term corporate bond could actually be sold to three
separate persons. One would supply the money for the bond; one would
bear the interest rate risk; and one would bear the risk of default.
The last two would not have to put up any capital for the bonds, although
they might have to post some sort of collateral.
-6-
The cost of transferring the bonds themselves, under these circumstances,
would be nearly zero. The buyer of the bonds would not have to worry
that the seller knew something about the company's fortunes, or about
future interest rates, that he did not know. The dealers in the bonds
would not be bearing any risks at all. The nominal transactions costs
could be paid by the buyers and sellers of the bonds.
In the same way, the function of a common stock in transferring risk
can be separated from its function as a store of wealth. When a company
sells stock, it would be possible for one person to put up the money,
and for another person to take on the risk. The first person would
simply be lending the second person the money to buy the stock. The
risk-bearer would normally give collateral to the lender, but he would
not have to put up any cash.
Again, the transfer of the risk would be independent of transfer of the
loan. The risk-bearer could transfer the risk to another person without
affecting the lender, so long as the new risk-bearer could post enough
collateral. And the lender could transfer the loan to another person
without affecting the risk-bearer.
The person who takes on the risk has to worry that the person giving
it up has inside information that he does not have. But a new lender
does not have to worry that the old lender has knowledge he doesn't
have. So long as the collateral is adequate, he is safe.
Thus the costs associated with transferring risky financial assets
should properly be associated with the transfer of risk, and not with
the transfer of funds.
CASH FLOWS AND FINANCIAL ASSETS
The two factors that are said to make an asset relatively illiquid
-7-
are risk and transfer costs. But we have seen that the risk of a
financial asset is separable from the asset as a store of wealth, and
that the transfer costs are associated with the risk of the asset
rather than with the asset as a store of wealth.
Thus liquidity in financial assets has nothing to do with cash flows.
Cash transfers are completely independent of risk transfers, and
liquidity, if it has any meaning at all, is associated with risk
transfers rather than with cash transfers.
One way to see this is to note that the purchase or sale of a financial
asset can be financed by a lender, just as the purchase or sale of a
real asset can be financed by a lender. The .time of the purchase or
sale is not related to the time at which the major portion of cash is
transferred.
If the seller provides credit to the buyer, then they can agree on a
mutually acceptable future time for payment to be made, and the cash
flow will be known in advance for both of them.
A person or institution does not need to hold liquid assets to help
handle unexpected cash flows. Almost all cash flows are expected,
and the cost and risk associated with handling those cash flows that
are unexpected are very low.
OPEN MARKET OPERATIONS
The idea that cash transfers are independent of risk transfers casts
new light on Federal Reserve open market operations.
-8-
The important difference between the Federal Reserve System and other
holders of government securities is that the Federal Reserve does not
choose its securities in order to maximize its income. Because of
this, the Federal Reserve actually loses money on its open market
operations; that is, it makes less money than it could make if it
held different securities.
The Federal Reserve is unique in that its decisions on what securities
to buy are not influenced primarily by expectations of future interest
rates. Since it is such a large factor in the government securities
markets, it can actually distort the term structure of interest
rates, so that government bond prices do not reflect only the market's
expectations of future interest rates.
To the extent that the Federal Reserve does not base its actions
on expectations of future short term interest rates, there are
profit opportunities for the other participants in the government
securities markets. What the Federal Reserve loses on its open
market operations, other participants, who trade with the Federal
Reserve, make.
It might seem that the Federal Reserve was in a position to make
money at the expense of othe~ participants, because of its great
knowledge of the market ruld its knowledge of its own intentions.
But in fact the reverse is true.
Since Federal Reserve actions create continuing profit opportunities,
the market is continually out of equilibrium. If the market were
more efficient, and transactions costs were lower, and knowledge of
Federal Reserve actions more widely available, the Federal Reserve
would not be able to influence market prices appreciably.
The inability of the Federal Reserve to make money from knowledge
of its own intentions depends on the fact that it is not able to
-9-
influence very short interest rates; in other words, that
it is able to influence futures prices but not spot prices.
In the very short end of the spectrum, quoted interest rates provide
all the knowledge an investor needs, and arbitrage is very effective.
Given a choice between competing instruments wihh the same risk
characteristics and the same maturity, the investor chooses the one
with the lowest price.
The only exception to this rule is in the case where "institutional
factors" prevent large holders of short term notes from choosing
certain classes of notes. Then there may be a significant differential
between the notes they can hold and the notes they can't hold, and
Federal Reserve actions may affect the size of this differential. But
this does not mean that the Federal Reserve is affecting "the" spot
rate of interest.
The Federal Government's budget will affect short term interest
rates. The larger the deficit, the higher interest rates will be,
as private investment is displaced by government borrowing. But
the actions of the Federal Reserve in shifting the composition of
the government's debt will not affect short term interest rates.
There is only one short term rate that is consistent with equilibrium
in the investment market, given the rate of change of prices. And
Cagan has shown that the price level cannot bring equilibrium to a
system in which it is an endogenous variable. So we must regard the
time path of prices as exogenous, and the short term interest rate
as endogenous. But in such a system, the actions of the Federal
Reserve are not necessary (or even possible) to determine the short
term interest rate.