Page 84 - Special Topic Session (STS) - Volume 4

STS566 Richard Finlay

Forecasting banknote demand at the Reserve
Bank of Australia
1
Richard Finlay
Reserve Bank of Australia

Abstract
I detail the Reserve Bank of Australia's method for forecasting banknote
demand and deciding on the level of contingency banknote stocks to hold. In
approaching this, the RBA has two main objectives: most importantly, to
always have sufficient banknotes to meet public demand; and second, subject
to the first requirement, to minimise cost. Regarding banknote forecasts, we
use statistical autoregressive models rather than models based on economic
and financial variables. There are two main reasons for this: autoregressive
models have historically performed reasonably well; and using explanatory
variables such GDP, the number of ATMs, etc. to forecast banknote demand
necessitates forecasting these variables, which in practice has been difficult to
do accurately. Autoregressive models will miss turning points or economic
shocks, however, and so we hold sufficient buffer stocks to allow for this and
ensure that we do not run out of banknotes. I also briefly discuss methods that
we have employed to estimate the proportion of banknotes in circulation used
for transactional and store-of-value purposes.

Keywords
bank note demand; forecasting; Central bank; contingency banknote stocks

1. Introduction
Banknotes, being complex physical objects, cannot be created instantly. In
fact, the sophistication of modern banknote designs and the number of
different components that they contain means that it can take many months
between notifying a printworks that more banknotes are required, and the
finished product being supplied to the central bank. Demand for banknotes,
on the other hand, can change very rapidly, as was the case in Australia and
many other countries around the onset of the global financial crisis. This miss-
match between a relatively slow production process and potentially fast
changes in demand, coupled with central banks’ unwillingness to countenance
running out of banknotes, means two things: central banks try to forecast
future banknote demand as accurately as possible; and, knowing that they will
nonetheless get things wrong, they employ backup strategies such as holding
substantial contingency stocks. This paper discusses the Reserve Bank of

1 The author is from Note Issue Department and would like to thank Ben Smagarinsky for his
input.
73 | I S I W S C 2 0 1 9