Page 96 - Special Topic Session (STS) - Volume 4

STS566 Iluminada T. Sicat
levels, 2) the level of currency stock necessary for safety cushion, and 3) the
desired fitness level of the banknotes and coins in circulation before they shall
be replaced. BSP’s own statistical tests showed that economic activity, as
captured by GDP, and the price level have long-run and strong co-integration
relationship with currency demand. There exists linear relationship between
demand for banknotes and these two variables. Specifically, increases in real
economic growth lead consumers to increase their usage of all denominations,
whereas an increase in price levels lead to an increase in the demand for higher
denomination bank notes. The use of GDP targets and price level expectations
provide the forward-looking information into the forecasts.

4. Currency Demand Forecasting Models
Forecasting model to estimate currency demand is assessed periodically
for robustness of the model. In this respect, the BSP used different forecasting
models over time to enhance the reliability of the forecasts under each model.
In the late 1990s, currency in circulation for the forecast year is derived from
the projected M3 based on the historical share of CiC to M3. Annual M3
projections, in turn, were based on a target real GDP growth and expected
inflation rate. Currency demand is then calculated based on change in CiC.

Model 1: (1999 – 2004) ( )
3

Starting 2005, CiC for the forecast year is projected by multiplying previous
year’s CiC by a growth factor using macro assumptions of real GDP growth
target and expected inflation rate.

Model 2: (2005 – 2-11) (CiC growth)t = inflation + 1.17 *(real GDP growth)t

By 2012, CiC for the forecast year is based on an OLS (ordinary least square)
5
model using consumer price index (CPI) and GDP as explanatory variables. A
dummy variable and an error-correction term are added to the baseline model
to capture the effects of global financial crisis on demand for banknotes, as
well as the difference between the forecast and the actual currency in
circulation. It was noted that financial market uncertainty and financial
volatility caused demand particularly for higher denomination banknotes to
rise.
Model 3: (2012 – 2017) LOG(CiC)t=Ii0 + Ii1 LOG(CPI)t + Ii2 LOG(GDP)t
+ Ii3DFIN08t + t

5 A linear regression model that aims to minimize the sum of the squared error.
85 | I S I W S C 2 0 1 9