Finding Indicators

Download Report

Transcript Finding Indicators

PFTAC GDP Compilation and
Forecasting Workshop
Forecasting components of
GDP
Suva, Fiji
October 20, 2016
Forecasting components of GDP
•
The first step is to find indicators.
•
Time series models can be developed
subsequently.
•
Know what the data is measuring
•
Always plot the data in levels and growth rates
•
… with leads and lags if it makes sense
•
… and moving averages
•
Plot the components of GDP as percent of total
GDP—the structure of the economy (e.g. the
relative size of industries for GDP(P) and of
sectors for (GDP(E)) usually only changes slowly.
2 of 8
Find stable relationships of the component
to forecast with other variables
•
Find external forecasts, e.g. consensus forecasts,
IMF, World Bank, population projections
The Pacific produces tuna, the IMF and World
Bank only forecast the price of salmon. Tuna and
salmon are substitutes—prices are expected to
move together over time.
World commodity prices lead import prices. It
takes time for imports to arrive in the Pacific.
World price of cotton may be a leading indicator
for textile prices.
3 of 8
Find stable relationships of the component
to forecast with other variables
•
Find leading indicators
Cement imports
Building permits
Private sector credit
Business sentiment
Job ads
Mining permits
Cruise ship table time
Weather patterns (southern oscillation index)
4 of 8
Decompose the series to forecast
•
Example: Exports of services
Exports of services mainly consist of tourism
spending.
Tourism spending = Number of tourist arrivals *
average number of days stayed * average daily
spending per tourist
Grow average daily spending per tourist by CPI
for nominal forecasts
For long-run forecasts: Use population
projections for Australia, New Zealand?
5 of 8
Decompose the series to forecast
•
Example from balance of payments: Remittances
of seasonal workers
Remittances = Number of seasonal workers *
hours worked per week * weeks worked * aftertax hourly wage
The number of seasonal workers is a function of
the total number of seasonal workers under the
scheme and the country’s uptake of the scheme.
6 of 8
Find patterns in the data
•
Stable ratios, shares
Example: Imports of services mainly consist of
transport and storage. Imports of services are
expected to be correlated with imports of goods.
•
Growth rate fluctuates around a constant level.
•
Ratios, shares, growth rates gradually return to
long-run averages over time.
7 of 8
Use the last period outcome
•
… if everything fails.
•
E.g. exchange rates, some components of the
balance of payments
•
Inflate for long-run nominal forecasts if needed.
8 of 8