Professional Documents
Culture Documents
The theoretical overview of the ECA macro model and its key theoretical foundations;
In line with ECA mandates to build State members capacity to design policy informed
by economic analysis
Good Forecasting and evaluation of various shocks and policies are critical to good policy
I. ECA Model: type & structure 5
The model consists of four sectors giving a comprehensive view of the entire economy
Households: consumption, savings, supply labor and capital
Firms: use of labour and capital to produce output
The government and the central bank: conduct fiscal policy and monetary policy
Foreign sector: record the balance of payments
Characterized by the long-run neoclassical supply-side and the short-run Keynesian demand side
Follows the specification of the world economic framework model (WEFM) and oxford Economics
Consumer spending is assumed to be consistent with the life cycle and permanent income
theories
Model Variables 7
Variables divided into demand and supply and into core and non-core variables
Core demand variables: aggregate expenditure components at constant and current prices;
monetary policy variables and financial variables
Core supply variables: prices, natural levels of output, unemployment and real wages
Non-core variables are specific to countries' objective and depending on data availability:
Inventory levels
Prices
Exports and Imports Exchange rate
Based on annual data, the sample period for model specification ends at 2016.
Simulation type:
Deterministic simulation: single forecast rather than a distribution of feasible values.
Stochastic simulation: distribution of forecast outcomes for each period in the forecast horizon
based on the random components of the model (the model's error term and its coefficient
uncertainty)
Static forecasting: use of actual rather than forecasted value for the lagged variable provided
that actual data are available.
Dynamic forecasting: use of previously forecasted values of the lagged left-hand variable to
forecast for the periods after the first period in the sample.
Solution scenarios:
Actuals: represent the original series
Baseline: serves as a reference point when forecasting
Solution sample: specify the sample period of interest.
In-sample vs Out of sample Forecasts 16
To do deterministic forecast, click on “solve” in the model window and specify in the dialog box as indicated above (left)
Compare the forecast variables ending with _0 with the real values:
Hold CTRL key; select variables of interest; right click and “open” as a “group”
Deterministic Forecasts Using the Model Object 2/3 18
Select as follows
To graph forecast variables: click on “proc” in the model window and select “make graph”
In the graph dialogue box, specify as indicated above (right visual aid)
In the graph window, the active scenario is termed “new”. It can be also the “baseline” there.
We will plot variables over the recent history from 2013. Let’s click “OK”
Deterministic Forecasts Using the Model Object 3/3 19
Fixed Capital Formation/GDP GDP growth Productivity per worker Household consumption
Actual values of the variables are in green while the model’s forecasts are in blue
Stochastic Forecasts Using the Model Object 1/3 20
Select as follows
Select as follows
“proc” “Make Graph”
Fixed Capital Formation/GDP GDP growth Productivity per worker Household consumption
Forecasted values in blue are within the confidence bands which take into account uncertainty
Alternative Forecasting Scenarios 23
Select
Inflation, cpi Public Debt Exports, goods and services Gross domestic product
It is recommended to assess forecast performance using out of sample forecast tests rather than in-
sample tests.
Small Root Mean Square of Errors (RMSE) is a good pointer for good forecast. Compute the RMSE
for every forecast horizon to distinguish between competing models
𝒇
United Nations Economic Commission for Africa (2019). Theoretical Foundation for the
Macroeconomic Model
Macroeconomic and Financial Management Institute of Eastern and Southern Africa (2012).
Macroeconomic Modeling and Forecasting Manual
United Nations Department of Economic and Social Affairs (2016). The World Economic Forecasting
Model at the United Nations
28
THANK YOU