BID BEHAVIOUR AND THE DETERMINATION OF UK TREASURY BILL RATES 1970–1976*
| Date | 01 August 1979 |
| Published date | 01 August 1979 |
| Author | CHARLES WARD,ANTHONY SAUNDERS |
| DOI | http://doi.org/10.1111/j.1468-0084.1979.mp41003003.x |
BID BEHAVIOUR AND THE DETERMINATION OF UK
TREASURY BILL RATES 1970_1976*
By ANTHONY SAUNDERS and CHARLES WARD
I. INTRODUCTION
In the United Kingdom treasury bills are allotted by means of a weekly
auction.' At this auction the Bank of England adopts the role of a price
discriminating monopolist, allocating bills in descending order of bid price until the
total supply is exhausted. Those bidders who tender at low prices (high) yields are
often disappointed, as the total demand is always greater than the supply.
The major competitors at this auction are: overseas holders of sterling, industrial
companies, non-bank financial intermediaries and a specialized group of money
market institutions, the discount houses. Most of these bidders have a great deal of
freedom in formulating and entering their bids. Even the discount houses, who
were traditionally constrained by a number of collusive agreements on their
tenders, now have a good deal of individual freedom in determining their bids.2
The aim of this paper is to analyse bidding behaviour and rate determination
in the UK treasury bill market over the period 1970-76. Initially two
representative models of bidding behaviour are discussed. The first is based on
rather naive assumptions about investors' behaviour, assuming a symmetric
rectangular (or uniform) distribution of bids. The second is based on a more
rational bidding strategy, which assumes that bids are asymmetrical and
exponentially distributed. Two important statistical characteristics of these
distributions are defined and empirically measured. In particular, it is argued that
since the 'treasury bill rate' currently published by the Bank of England [2] is based
only on the average of accepted bids, it is unrepresentative of the whole distribution
of bids. Hence its use in empirical studies may result in a loss of important
information, especially on the nature and determination of bids at the lower end of
the distribution.
Using derived characteristics of the two bid distributions as dependent
variables a small number of independent variables are then identified. Tests arc
conducted to examine the extent to which bids in the bill market can be explained
within the framework of a relatively simple interest-rate determination model. It
is argued that while such models appear to offer some explanation of bidding
behaviour, investors do appear to be heterogeneous in their sensitivity to
determining variables. Moreover the more rational model (the exponential)
* We should like to acknowledge the helpful comments of P. Blandon, City of London Polytechnic
and A. Kalay and J. Kallberg, New York University, Graduate School of Business Administration.
I All new Treasury bills have 91 day terms to maturity and are sold in six denominations ranging
from £5,000 to £250,000. The minimum size of any lot is £50,000.
2 The introduction of the Competition and Credit Control reforms of May-June 1971 resulted in the
abandonment of the houses collusive agreement whereby all bid at one price, collectively determined
prior to the auction, and their traditional obligation to cover the whole weekly supply of new bills was
replaced with a new, more flexible, agreement.
215
To continue reading
Request your trial