Upload
e-forecastingcom
View
373
Download
2
Embed Size (px)
DESCRIPTION
In his presentation, The Importance of Economic Conditions When Building Forecast Models, Chief Economist of e-forecasting.com Dr Evangelos Simos covers a wide variety of key economic concepts and the possible direct and indirect impacts each has on forecast accuracy. Although several forecast packages and methodologies exist, we all operate in the same economic environment and the changing conditions impact demand. Even if current corporate forecast processes do not take into account economic conditions, Dr. Simos' presentation will leave the audience with basic guidelines of how to interpret key economic events and their likely outcome on business. For instance, what are the implications of oil prices spiking to $150 this summer? How much worse are conditions expected to get in Europe and how will that impact exports and foreign demand? Will further Mid East tensions bring violence and political instability? What happens in the US if Obama is re-elected? e-forecasting.com, an international economic research and consulting firm, offers forecasts of the economic environment using proprietary, real-time economic indicators to produce customized solutions for what’s next. e-forecasting.com collaborates with domestic and international clients and publications to provide timely economic content for use as predictive intelligence to strengthen its' clients competitive advantage.
Citation preview
April 10, 2023 2
Indicators, Predictive Analytics & Forecasting
Evangelos Otto Simos, Ph.D.
Economic Research
e-forecasting.com
April 10, 2023 3
Outline
• Going Beyond Raw Data • Forecasting with Hundreds of
Drivers• Case: Forecasting a Hotel
Market• US & Global Outlook
April 10, 2023 4
GoingBeyondRawData
April 10, 2023 5
Types of Indicators
• Quantitative• GDP, Industrial Production, Employment
• Qualitative • NAPM Survey, Consumer Surveys
• Analytics (Diffusion Indices) • Aggregates of Quantitative and/or Qualitative
Indicators
• Predictive Analytics• Fact-based forward-looking analytics (composite
leading indicators)
April 10, 2023 6
Case: PMI: NAPM Diffusion Index
NAPM History
20.0
30.0
40.0
50.0
60.0
70.0
80.0
Jan-59 Jan-64 Jan-69 Jan-74 Jan-79 Jan-84 Jan-89 Jan-94 Jan-99 Jan-04 Jan-09
April 10, 2023 7
A closer Look at PMI Analytic
Closer Look: NAPM 2012-13
48.0
49.0
50.0
51.0
52.0
53.0
54.0
55.0
56.0
Jan-12 Apr-12 Jul-12 Oct-12 Jan-13
• Last three months of 2012 average: 50.6
• Manufacturing was stagnant in the last quarter of 2012
New Orders % Better % Same % Worse Net Index
Jan 2013 28 51 21 +7 53.3
Dec 2012 24 45 31 -7 49.7
Nov 2012 26 43 31 -5 51.1
Oct 2012 24 47 29 -5 52.8
April 10, 2023 8
Housing Sales: Is that a Recovery?
April 10, 2023 9
Consumer Confidence
• Media pays attention to composite indices• Consumer spending is about 70% of GDP• Overall indices found to coincide even to lag
overall economic activity (GDP)• Q: Are they useless predictive analytics?• A: Not at all• Why?
April 10, 2023 10
Consumer Confidence (continued)
• Consumers are not forecasters • Overall indices are loaded with answers on questions
about the economy • Look at answers to survey’s questions related to
individual consumer activity, i.e. about them • Are you going to buy a car over the next six months?
New? Used? etc• Are you going to buy a house over the next six
months? New, Existing? etc• Are you going to travel over the next six months?
Domestic? Foreign? By Plane? By Car?
April 10, 2023 11
Economic Policy: Fed in a printing mode again
April 10, 2023 12
Nominal Vs. Real
• Retail Sales• Construction Spending• Orders (for Durables, Capital Goods)• Average Wage Rate ($ per hour)• Personal Income• Exports
April 10, 2023 13
Dealing With Information
A modern data infrastructure
• Internal vs. external
• Timely and effective collection or updating tools
• Well-designed mining software
• Analytics processes to tap into meaningful & useful data for predictive intelligence
Analytics for predictive intelligence
• Eliminate transitory outliers, seasonally adjust and smooth out noise
• Identify hidden trends in economic, business and financial patterns
• Refine large information into a small set of market-centric composite predictors
April 10, 2023 14
Many Drivers: Blessing or Curse?
• Advances in information technology provide access to thousands of economic and business indicators
• In forecasting, "having many time series is a blessing not a curse”, James Stock1
• A new frontier in forecasting proposes to pool the information of all available predictors
1Department of Economics, Harvard University
April 10, 2023 15
Modeling Hundreds of Predictors
• Classical VS. Modern Modeling • Let y denote a business indicator to be
forecast (like Hotel Occupancy)• Let X be a small number of typical or
important variables that “classical” forecasters use as predictors, given statistical limitations (like GDP, Inflation, Unemployment)
April 10, 2023 16
Forecasting with Hundreds of Predictors• Let Z be a set of a few “composite” or so-called
“diffusion” indicators that capture a large number (several hundreds) of individual predictors, which • are individually unimportant • collectively become important• provide useful “missed” information, when
properly grouped• their “combined” contribution to predicting y
may be as good as or better than the typical set of predictors X
April 10, 2023 17
The Modern Approach
• The General Modern Econometric Model is y = αX + βZ + ε • Origins of the methodology
• Burns and Mitchell (1947) in studying business cycle indicators with composites
• Sargent and Sims (1977) on factor analysis• Stock and Watson (1989, 2004) on" forecasting using
many predictors”• Bernanke and Boivin (2002) on monetary policy in a
data-rich environment
1In their research works Forecasting using diffusion indexes, and Forecasting using many predictors.
April 10, 2023 18
Predictive Intelligence Modeling
Forecasting Model
IndustryMetrics
CompositeEngine
CompositeDrivers
Foreign MetricsNational
Metrics
RegionalMetrics
April 10, 2023 19
Case: Forecasting Hotel Market
• History • Deloitte (London), Hotel Benchmark
Division• STR Global• TRI & e-forecasting
April 10, 2023 20
Who Visits a Major City (London)?And, Stays in a Hotel
London
April 10, 2023 21
Purpose of Trip: Why you visit a city? And, Stay in a Hotel
Tourists Business Government Affairs
April 10, 2023 22
Where do Visitors come from? Domestic &
International Feeders to a Destination (Milan, Italy)
Origins of visitors – who stays in hotels - in a typical
destination city: Milan
• 41 percent of all visitors – 3 million - come from Italy (domestic feeder)
• 59 percent, or 4.3 million visitors come from the rest of the world• Majority of foreign visitors (76 percent) come from 31 countries.
Sample of some feeding countries:• 7.6% from the United Kingdom • 7.2% from the United States• 6.7% from Germany• 5.7% from France• 1.8% from Sweden• 0.6% from Mexico
April 10, 2023 23
What Drives Determine Visitors Volume for a Destination?
• Drivers for Tourism-Related Visitors• Personal Income, Employment Stability, Consumer
Confidence, Inflation, Capital Gains/Losses like Stock & Housing Prices, Interest Rates, Value of Domestic Currency (Exchange Rates) etc
• Drivers for Business-Related Visitors• Measures of Business Sentiment (Future production,
Sales, Incoming Orders), Investors Financial Optimism, Profitability, Interest Rates, Taxation, Energy costs, etc
• Drivers for Government-Related Visitors• Measures Affecting Revenues (Income, Employment
and Sales Taxes), Cyclical Spending (Unemployment), Public Investment, etc
April 10, 2023 24
Forecasting with Hundreds of Drivers?
The Challenge • Multifaceted Drivers
• Domestic
• Many countries (Top 30)
• Type of visitors (three)
• Indicators per dimension (about 10)
• The Upshot: Hundreds of Drivers
15GermanyDrivers
15India
Drivers
15 Brazil Drivers
15USA
Drivers
20DomesticDrivers
Hotel Market
April 10, 2023 25
Case: Modeling London Hotel Market• Modeling Approach
• Refining large information (indicators) into a small set of composite drivers (predictors) for the London hotel market
• Drivers are observable economic and business composites that influence business activity in London’s hotel market
• Occupancy (OCCU) and room rates (ADR) are interrelated
• Domestic and international drivers influence demand for and supply (including costs) of hotel services to determine future hotel performance
April 10, 2023 26
Adjustment of Hotel Market Indicators• To identify the “best fit” and generate most
accurate forecasts, hotel-market indicators are brought to conformity with single drivers and composites, which are seasonally adjusted by source or e−forecasting.com
• Hotel market indicators are adjusted for seasonality and special “transitory” events such as Royal weddings, sport events, political events, etc. Events are also treated in the future for prediction of performance metrics.
April 10, 2023 27
Composite Drivers
• Foreign Composites: Combine activity, current and future,
from all countries that feed London according to their relative importance (grouping of more than 1,000 indicators)
• Business expectations• Consumer confidence• Incomes• Changing value of assets• Exchange rates
• Domestic Composites: Combine activity, current and future, in the domestic economy
• Consumers• Business• Financial markets• Costs (wages, prices, energy)
April 10, 2023 28
London Forecasts: Occupancy & Room Rate
London Hotel Market: Occupancy Forecast
55
65
75
85
95
01 02 03 04 05 06
Forecast of Baseline Actual Baseline, SA&S
Forecast of Occupancy Actual Occupancy
Room Rate Forecast
£85
£95
£105
£115
£125
£135
£145
01 02 03 04 05 06
Forecast of Baseline Actual Baseline, SA&S
Forecast of Room Rate Actual Room Rate
April 10, 2023 29
US Economic Outlook
• Measuring long-term economic performance with stud metrics
• USA 1980-2007 model: 4x2• 4% growth & 2% Inflation
• European 1980-2007 model: 2x2• 2% growth & 2% Inflation
• Forecast: USA new reality: 2x2 or 2x4 ?
April 10, 2023 30
Growth in Monthly GDP
• Looking @ growth rate, can see depth of previous recession, upswing, and now slow recovery in economy
• When negative, recession; when positive, expansion
• Six-month growth rate, which signals confirmation of turning points, went up 1.2% in January, after going up 1.4% in December
30
April 10, 2023 31
Current Risks
• Ineffectiveness of low interest rates
• Asset inflation
• Deficits, Debts, Uncertainty and Fear – The European experience has moved west
• Free fall of dollar, panic, policy reversal, high interest rates, a “real” depression
• Geopolitical factors, Middle East conflict & oil prices
April 10, 2023 32
Dynamic Sector: US Manufacturing
Manufacturing recovery will slowly return to peak of 08 by 2015
April 10, 2023 33
Monetary Policy on the Horizon
• Look for the Fed Funds rate to remain near zero until 2015
• Bernanke’s ‘pledge’ to hold rates thru 2014
3-Month Treasury Bill: Secondary Market
-2
0
2
4
6
8
03:Jan 05:Jan 07:Jan 09:Jan 11:Jan 13:Jan 15:Jan
Percent
April 10, 2023 34
Global Outlook
• Looking @ major economic blocs and their leading indicators helps give an idea of turning points, which areas suffered more than others and which are recovering…
April 10, 2023 35
Long-Term Global Outlook
• Market Size Measured by GDP in $PPP Billion in 2011
April 10, 2023 36
Emerging Asia shows continued dominance in global economy
CONTRIBUTION OF REGIONS TO GLOBAL GROWTH
REGIONPercentage Points Contribution Relative Contribution, Percent
2011 2012 2013 2014 2011 2012 2013 2014
EUROPEAN UNION (EU27) 0.35 -0.04 0.02 0.26 9.4 -1.4 0.7 8.1
Euro Area (euro17) 0.23 -0.07 -0.05 0.13 6.1 -2.5 -1.8 4.1
Non-Euro Members (10) 0.12 0.03 0.06 0.13 3.3 1.1 2.5 4.0
OTHER EUROPE 0.30 0.17 0.20 0.24 8.0 6.1 7.9 7.4
NORTH AMERICA 0.52 0.52 0.32 0.54 13.8 18.4 12.5 17.1
United States 0.38 0.39 0.22 0.42 10.1 14.0 8.9 13.3
SOUTH AMERICA 0.28 0.17 0.20 0.24 7.5 6.0 7.9 7.4
ASIA & PACIFIC INDUSTRIAL 0.06 0.19 0.14 0.19 1.7 6.6 5.5 5.8
EMERGING ASIA 2.03 1.67 1.52 1.56 54.1 59.4 59.7 48.9
China & India 1.76 1.42 1.26 1.28 46.8 50.5 49.5 40.1
MIDDLE EAST & AFRICA 0.15 0.09 0.10 0.13 3.9 3.1 3.9 4.2
WORLD GROWTH1 3.8 2.8 2.5 3.2 100.0 100.0 100.0 100.0
1Sum of Regional Contributions Source: www.e-forecasting.com
April 10, 2023 37
Short term global forecast% change in real GDP growth
-1012345678
EUR
O A
REA
(17)
NO
N-E
UR
OAR
EA (1
0)
OTH
EREU
RO
PE
NO
RTH
AMER
ICA
SOU
THAM
ERIC
A
ASIA
&PA
CIF
ICIN
DU
STR
IAL
EMER
GIN
GAS
IA
MID
DLE
EAST
&AF
RIC
A
2011
2012
2013
2014
April 10, 2023 38
Long-term global forecast
-1
0
1
2
3
4
5
6
7
8
2011
2013
2015
2017
2019
2021
2023
2025
2027
2029
EURO AREA (16)
NON-EURO AREA (11)
OTHER EUROPE
NORTH AMERICA
SOUTH AMERICA
ASIA & PACIFIC INDUSTRIAL
EMERGING ASIA
MIDDLE EAST & AFRICA
April 10, 2023 39
Q&A
April 10, 2023 40
Thank You
For questions related to this presentation please contact