1. PRICING & PURCHASING SUMMIT
WASHINGTON DC | NOVEMBER
PRICING & PURCHASING SUMMIT

2. Confidence Intervals with IHS Forecast
Training Track
Confidence Intervals with IHS Forecast
11, November 2014
KC Chang
Senior Economist

3. Objective Forecasting is a challenging task
Pricing & Purchasing Summit / November 2014
Objective
Forecasting is a challenging task
Making critical business decisions with forecast data is also a difficult task
We have developed a new capability to show you
It shows a forecast as being a range of possibilities
There are other ways to forecast and analyze risk
We want your feedback.

4. Pricing & Purchasing Summit / November 2014
Overview
The IHS Pricing and Purchasing (P&P) service provides 10-year price forecast for numerous raw commodities, semi-finished and finished products.
The forecasts are point estimates. It is an expected value or average price for a specific point in time.
Point estimates form the IHS P&P baseline price forecast. The estimates are specific but struggle to address the other risk of forecasting…volatility.
Procurement executives and project managers often face a environment where prices for crucial materials fluctuate over a specified time period.
The level of price volatility affects the timing of purchases and key capital expenditure project approval.

5. Historical Price Volatility: LME Copper Spot Price
Pricing & Purchasing Summit / November 2014
Historical Price Volatility: LME Copper Spot Price

6. Uncertainty in Forecasts
Pricing & Purchasing Summit / November 2014
Uncertainty in Forecasts
The baseline price forecast assumes a ‘Constant World’.
The estimated economic relationships are fixed. The relationship does not change over time
The expected value is the most likely and it would take a large disruption to alter the forecasted outcome
If there is a disruption, it is very unlikely. The chances of a forecast ‘overshoot’ is the same as a forecast ‘undershoot’
In reality,
The daily, weekly, monthly spot price can exhibit strong fluctuations. For some commodities, the econometric relationship may have a stronger upward/downward bias influencing the forecast.
Project estimation costs can escalate well-beyond expectation even when there is no significant change to the economic outlook.

7. Static Confidence Intervals
Pricing & Purchasing Summit / November 2014
Static Confidence Intervals
2016Q1: $15,000
2016Q1: $-2000
The conventional method to visualize forecast uncertainty is using confidence intervals for the forecast values.
This is normally an embedded feature for most economic software

8. Static Confidence Intervals
Pricing & Purchasing Summit / November 2014
Static Confidence Intervals
Advantage:
It is convenient. A quick estimate on the potential range of values
Drawbacks:
The interval is so large it provides minimal value. (Ex: the 95% confidence interval for copper prices showed potentially negative prices.)
Within the large interval, it is still difficult to visualize risks near the baseline forecast. (ie: The baseline is most likely, but are there also likely potential price values?)
It is challenging to decide if small changes to the baseline forecast signal more risk about the future. (ie: Is the recent price spike/drop an unusual event or signal of something else?)

9. Dynamic Confidence Intervals
Pricing & Purchasing Summit / November 2014
Dynamic Confidence Intervals
It is a graphing method to convey the risk around a forecast.
Forecast comes from using a bootstrap method to estimate economic relationship.
We create many forecasts through (ie: 1000, 10000, etc.) simulation. This generates a whole distribution of forecast price outcomes.
We create confidence intervals by looking at the distribution of prices within each time period. This places the baseline forecast in a larger context.
Benefits:
Forecast uncertainty is model-driven. Forecast error terms come from the forecast sample or it can be specified by the analyst
Helps integrate incoming market intelligence when analyzing the near-term outlook. Judgment is subjective but less ad-hoc.
Effective visual communication about potential risk to the outlook when holding the macroeconomic outlook fixed

10. Example: Dynamic Confidence Intervals
Pricing & Purchasing Summit / November 2014
Example: Dynamic Confidence Intervals
2016 Q1:$8050
2016 Q1: $5230

11. Application: Visualize Volatility
Pricing & Purchasing Summit / November 2014
Application: Visualize Volatility
For upstream commodities, a way to visualize market volatility and mitigate risks.
For procurement executives:
Looking at the upside and downside price within each forecast quarter/year helps buyer’s hedge risk.
The forecast simulation process is systematic but flexible. Users can specify the error terms to see the impact of potential non-macroeconomic shocks.
Clients can use own historical data to generate the randomness affecting the baseline forecast. OR specify new simulations to test alternative planning scenarios.
Users can also exogenously shock the economic variables driving the baseline price forecast. After the exogenous shock, an alternative baseline and confidence interval chart can be produced.

12. Dynamic Confidence Intervals for Cost Analyzer
Pricing & Purchasing Summit / November 2014
Dynamic Confidence Intervals for Cost Analyzer
The Cost Analyzer is the standard tool accessed through Data-Insight web.
Helps buyers analyze single commodity buys.
Helps buyers analyze costs for key capital cost projects
Helps buyers evaluate supply chain performance and cost benchmark key spend-categories.
Capability:
Confidence intervals can be applied to client-built composite cost indices.
It is possible to see how price volatility filters through the supply-chain.
Help set price escalation ranges with suppliers during contract negotiation

13. Example: Underlying Indices
Pricing & Purchasing Summit / November 2014
Example: Underlying Indices
Labor
Iron and Steel

14. Example: Underlying Indices Continued
Pricing & Purchasing Summit / November 2014
Example: Underlying Indices Continued
Energy
Using the three cost indices, I create a composite index:
COMPOSITE Index =
0.4*Labor +
0.4*Energy
0.2*Iron&Steel

15. Example: Final Composite Index
Pricing & Purchasing Summit / November 2014
Example: Final Composite Index

16. Application 2: Project Planning
Pricing & Purchasing Summit / November 2014
Application 2: Project Planning
Enhancement of the Cost Analyzer for Cost Estimators and Strategic Planners
Benefits:
Large capital expenditure project can experience stronger than expected cost escalation from non-macroeconomic factors (ie: geographic location, industry specific risks, unexpected delays/events).
Clients can use own historical data to generate the randomness affecting the cost analyzer forecast. OR specify new simulations to test alternative planning scenarios.
Effective visual communication about potential project when discussing investment decisions with interval departments and senior management.

17. Next Steps: The even bigger picture
Pricing & Purchasing Summit / November 2014
Next Steps: The even bigger picture
The dynamic forecast charts display uncertainty around the baseline forecast
Scenario analysis is the other technique we use to analyze risk and explore the range of potential outcomes
We are working on towards connecting all IHS economic models
Simulations can be connected to scenarios

18. A cutting-edge economic model that links global interdependent markets
Pricing & Purchasing Summit / November 2014
A cutting-edge economic model that links global interdependent markets

19. Global Link Model: Encompasses 95% of global GDP
9/20/2018
Pricing & Purchasing Summit / November 2014
Global Link Model: Encompasses 95% of global GDP 68
country models...
Europe:
Austria, Belgium, Bulgaria, Croatia, Cyprus, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, Netherlands, Norway, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, Sweden, Switzerland, Russia, Turkey, UK
North America:
Canada, Mexico, US
Africa:
Algeria, Angola, Egypt, Morocco, Nigeria, South Africa, Tunisia
Middle East:
Iran, Israel, Kuwait, Qatar, Saudi Arabia, UAE
Latin America:
Argentina, Brazil, Chile, Colombia, Peru, Venezuela
Asia-Oceania:
Australia, China, Hong Kong, India, Indonesia, Japan, Malaysia, New Zealand, Philippines, Singapore, South Korea, Taiwan, Thailand, Vietnam

20. Global Link Model: Incorporates key economic drivers
9/20/2018
Pricing & Purchasing Summit / November 2014
Global Link Model: Incorporates key economic drivers
...linked with each other and with global drivers of change
CAPITAL FLOWS
FDI allocation
foreign reserve allocation
new credits & debt financing
Europe:
Austria, Belgium, Bulgaria, Croatia, Cyprus, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, Netherlands, Norway, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, Sweden, Switzerland, Russia, Turkey, UK
North America:
Canada, Mexico, US
COMMODITIES
Agricultural:
Wheat, rice, corn, cotton, soybeans, cocoa, coffee, sugar, vegetable oils
Non-Agricultural:
Aluminium, copper, nickel, tin, zinc, iron ore, gold
TRADE FLOWS
Agricultural commodities
non-agricultural commodities,
manufactured products, energy
Africa:
Algeria, Angola, Egypt, Morocco, Nigeria, South Africa, Tunisia
Middle East:
Iran, Israel, Kuwait, Qatar, Saudi Arabia, UAE
Need a new base map and a design for the regional ‘pods’ of countries covered
Latin America:
Argentina, Brazil, Chile, Colombia, Peru, Venezuela
Asia-Oceania:
Australia, China, Hong Kong, India, Indonesia, Japan, Malaysia, New Zealand, Philippines, Singapore, South Korea, Taiwan, Thailand, Vietnam
ENERGY PRICES
Oil prices, gas prices, coal prices, electricity prices 

21. Commodity Prices Singled Out
Pricing & Purchasing Summit / November 2014
Commodity Prices Singled Out
Agricultural commodities
Wheat
Rice
Corn
Cotton
Soybeans
Cocoa
Coffee
Vegetable Oils
Non-agricultural commodities
Aluminum
Copper
Nickel
Tin
Zinc
Iron Ore
Gold
Energy
Oil prices
Brent
WTI
Gas prices
Europe
North America
Rest of World
Coal prices
Atlantic
Pacific
68 electricity prices (retail prices)

22. Global Link Model: Leverages comprehensive data
9/20/2018
Pricing & Purchasing Summit / November 2014
Global Link Model: Leverages comprehensive data
time series per country
Output & retail
Energy
Trade Statistics
Population, labor & wages
Monetary & financial
Prices & deflators
National income account
Household finance
Tax rates
Government finance
Exchange rate
Balance of payments

23. Example GLM Scenarios Fiscal policy scenarios Commodity price shocks
Pricing & Purchasing Summit / November 2014
Example GLM Scenarios
Fiscal policy scenarios
Changes in different tax rates
Changes in overall budget policy stance
Impact of environmental policies, carbon pricing
Commodity price shocks
Oil, gas, coal price shocks
Commodity prices shocks
Scarcity scenarios
Other exogenous shocks
Olympics, World Cup
Geopolitical stress
Trade wars
Other supply shocks
Monetary policy / Financial shocks
Changes in interest rates
Unconventional monetary policy measures
Changes in financial market stress, and/or in the degree of aversion to risk
Asset price shocks
Debt sustainability
Demographic changes
Changes in demographic patterns / migration flows
Pension reform
Changes in labor force participation rates
Changes in urbanisation rates
Changes in income distribution
Exchange rate shocks
Currency wars
Risk induced capital flight
Change in external reserves’ and portfolio allocation
New exchange rate regimes (emerging economies)

24. Pricing & Purchasing Summit / November 2014
Summary
Dynamic Confidence Intervals are a new capability for Pricing and Purchasing subscribers
It helps clients better understand uncertainty and we think it can help mitigate risks throughout the supply chain and project planning process
Please contact your Account Manager if are you interested in learning more about this new capability

25. Thank You!
Questions?