Required ResourceTextSchneider, A. (2017). Managerial Accounti.docx
Forecasting the Recession
1. Canada’s Financial Future
The Canadian Business Cycle
Created by Kirsten Boer, Shannon Hamilton & Gretl
2. Agenda
1. Overview of Topic
2. Economic Theory
3. Data & Methods
4. Empirical Results & Sensitivity Analysis
5. Practical Applications
6. Summary
3. Overview of Topic
What is the future path of Canada’s
Business Cycle? Is Canada’s
economy currently contracting?
Overview Economic Theory Data & Methods Results Application Summary
4. In the News
OECD warns of worst recession since early 1980s
OECD says developed world could face worst recession since early 1980s; warns of
deflation
Economic contraction in Canada will be short-lived:
Royal Bank
U.S. economic woes not necessarily Canada's fate
Overview Economic Theory Data & Methods Results Application Summary
5. Business Cycles
Business cycle:
“fluctuations of
Peak
aggregate economic
activity”, which
consists of 4 stages:
contraction, trough,
expansion & peak Trough
Overview Economic Theory Data & Methods Results Application Summary
6. Business Cycles: Burns & Mitchell
Factors
Business Cycles that affect
are recurrent, but not business cycles have
periodic regular patterns of
behavior
Theory
suggests an
ARIMA model
would not be
suitable
Overview Economic Theory Data & Methods Results Application Summary
7. ARIMA & GDP’s Growth Path
Overview Economic Theory Data & Methods Results Application Summary
8. Classes of Economic Variables
Leading Coincident Lagging
Variable’s turning Variable’s turning Variable’s turning
point occurs before point occurs at point occurs after
the business cycle the same time as the business cycle
the business cycle
Theory suggests the appropriate lag length of different variables
Overview Economic Theory Data & Methods Results Application Summary
9. Classes of Economic Variables
Procyclical Acyclical Countercyclical
Variable does not
Variable moves in Variable moves in
move with the
the same direction the opposite
business cycle
as the business direction as the
(e.g. real interest
cycle business cycle
rates)
Theory suggests the coefficient sign of different variables
Overview Economic Theory Data & Methods Results Application Summary
10. Considered Variables
•Durable consumption •Government Spending
• Stock index prices on G&S
GDP= C + I + G + NX
•Capital Investment • Trade Balance
• CA Balance
•Exchange rates
Other considered Variables: Bank of Canada Interest rate, Labour productivity, inflation
(CPI), unemployment, MB1, mortgage rates
Overview Economic Theory Data & Methods Results Application Summary
11. What period should be used?
• 1996:1 to 2006:1
• Allows for a hold-out period
• Uses data that best matches the current structural
relationship between the independent variables and
the dependent variable
Overview Economic Theory Data & Methods Results Application Summary
12. Is the Data Stationary?
GDP Labour Productivity
Industrial Production Capital Investment
Overview Economic Theory Data & Methods Results Application Summary
13. Is the Data Stationary?
Exchange Rate
Overview Economic Theory Data & Methods Results Application Summary
14. Correlation Matrix
Correlation Coefficients, using the observations 1996:1 - 2006:1
d_lab d_industprod d_exchange d_kinvest
1.000 0.3725 -0.0107 -0.1914 d_lab
1.000 -0.1136 -0.0106 d_industprod
1.000 0.1563 d_exchange
1.000 d_kinvest
Overview Economic Theory Data & Methods Results Application Summary
15. Chosen Model
coefficient std. error t-ratio p-value
--------------------------------------------------------------
const 3382.92 928.731 3.643 0.0010 ***
d_lab 4419.39 1049.81 4.210 0.0002 ***
d_industprod_1 0.524218 0.0982486 5.336 9.03E-06 ***
d_Exchange_1 !797.298 213.312 !3.738 0.0008 ***
d_kInvest_1 0.302776 0.0367233 8.245 3.34E-09 ***
Mean of dependent variable = 9501.31
Standard deviation of dep. var. = 5024.55
Sum of squared residuals = 1.19761e+08
Standard error of the regression = 1965.52
Unadjusted R-squared = 0.86446
Adjusted R-squared = 0.84698
F-statistic (4, 31) = 93.4873 (p-value < 0.00001)
Overview Economic Theory Data & Methods Results Application Summary
16. Review of Parameter Signs
Variable Sign Reasoning
Labour is better utilized in a period of economic
d_labour + prosperity
d_industrialprod_1 + Firms produce more when the economic outlook is good
Canada is a net export country; as the change in the
d_exchange_1 - exchange rate falls, Canadian exports become more
attractive
Firms invest more when the economic outlook of
d_kinvestment_1 + Canada is good
Overview Economic Theory Data & Methods Results Application Summary
17. Specification Tests
Test Output Result
Ramsey RESET p-value=0.2706 DNR Null Hypothesis
White Test p-value=0.3597 DNR Null Hypothesis
Breusch-Godfrey p-value=0.166879 DNR Null Hypothesis
Graph does not cross CI
CUSUM/CUSUMSQ No structural Break
Bands
Overview Economic Theory Data & Methods Results Application Summary
18. In-Sample Forecast Accuracy
Measure Value
MFE 0.000625
MAFE 1408.3875
MSE 2678972.259
MPE 0.015954892
MAPE 0.1514023
Theil’s U 0.078145808
Overview Economic Theory Data & Methods Results Application Summary
22. Sensitivity Analysis
1. What if the future values of the independent
variables were less volatile than expected?
2. What if government successfully increase capital
investment?
Overview Economic Theory Data & Methods Results Application Summary
23. Sensitivity Analysis
Comparison of Different Change in GDP Predictions
30000
22500
15000
7500
0
-7500
-15000
-22500
-30000
2001:02:00 2002:01:00 2002:04:00 2003:03:00 2004:02:00 2005:01:00 2005:04:00 2006:03:00 2007:02:00 2008:01:00 2008:04:00 2009:03:00
d_GDP Initial Prediction
Alternate Prediction (1) Alternate Prediction (2)
Overview Economic Theory Data & Methods Results Application Summary
24. Independent Variables & Theory
Independent Variables Theory Suggests
1. Labour Productivity 1. Leading/ Procyclical
2. Industrial Production 2. Coincident/ Procyclical
3. Exchange Rate 3. No theoretical view
4. Capital Investment 4. Coincident/ Procyclical
Overview Economic Theory Data & Methods Results Application Summary
25. Independent Variables & Theory
Independent Variables Model Suggests
1. Labour Productivity 1. Coincident/ Procyclical
2. Industrial Production 2. Lagging/ Procyclical
3. Exchange Rate 3.Lagging/Countercyclical
4. Capital Investment 4. Lagging/ Procyclical
Overview Economic Theory Data & Methods Results Application Summary
26. ARIMA Forecasting Accuracy
Measure Value
MFE -2,161.71
MAFE 3,538.348
MSE 17,710,115
MPE -0.318
MAPE 0.425088
Theil’s U 0.163985
Overview Economic Theory Data & Methods Results Application Summary
27. ARIMA vs. OLS Model
Our
ARIMA findings match
provides a less Burns &
accurate model Mitchell’s
Theory
Overview Economic Theory Data & Methods Results Application Summary
28. Macro-Level Application
Framework to Direct Timeline
Create Public Estimations
Government’s
Policy
Public Policy
Analysis &
Comparison Tool
Overview Economic Theory Data & Methods Results Application Summary
29. Micro-Level Application
Trend Economic
Analysis Decision
Criteria
The Graduate
Decision: School
or Work? Sensitivity
Forecasted Analysis
GDP values vs.
Market Beliefs
Overview Economic Theory Data & Methods Results Application Summary
30. Key Takeaways
1. GDP should begin to recover by the end
of 2008
2. Government should work to create
policies that aid in the recovery of
Canada’s economy
Overview Economic Theory Data & Methods Results Application Summary