1 / 31

Real-Time Forecasts for the World Economy: A Practitioner's Perspective

Real-Time Forecasts for the World Economy: A Practitioner's Perspective. Robin Koepke and Emre Tiftik Global Macroeconomic Analysis Department Institute of International Finance George Washington University Washington, DC April 26, 2012. 3 Key Messages about our Forecasts.

fauve
Download Presentation

Real-Time Forecasts for the World Economy: A Practitioner's Perspective

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Real-Time Forecasts for the World Economy:A Practitioner's Perspective Robin Koepke and Emre Tiftik Global Macroeconomic Analysis Department Institute of International Finance George Washington University Washington, DC April 26, 2012

  2. 3 Key Messages about our Forecasts We use few models and lots of judgment. Knowing the data is key. Forecasts are more than just numbers.

  3. Our Work We produce a monthly “World Economic Framework” with forecasts for all major global macro variables Forecasts include G3 growth, inflation, unemployment, policy interest rates etc. We have a small team of 4-5 people, who devote about 20-30% of their time to the forecasts

  4. What’s the Purpose? Practitioner Forecasts Academic Forecasts Express a view Provide an anchor Tell stories Be right ("forecast accuracy") Find “truth”: understand & establish the relationships between variables …?

  5. Where’s the Model?

  6. Where’s the Model? Appropriate Time of First Policy Rate Hike as per FOMC Number of FOMC participants in favor of time shown on x-axis

  7. Then How Do We do Forecasts? • Short term (0-6 months): tracking data • Medium term (3-12 months): reconcile ST changes with LT view • Long term (6-24 months): adopt a view

  8. Why Use Judgmental Forecasts? (1) • Judgmental forecasts can use information that is hard to quantify • Do you really trust your model? (I only trust it if I like the result.) • Stories matter! • Consistency matters • Time is scarce

  9. Using Judgment: “Known Unknowns” United States: Fiscal Policy in 2013 Percent of GDP, calendar year basis

  10. Our Current U.S. Growth Forecast United States: Real GDP Contributions to growth in percentage points, saar

  11. Using Judgment: “Unknown Unknowns” Japan: Industrial Production and Real Exports Index, February 2011 = 100

  12. Why Not Use a Model? (1) • Last week’s presentation: oil price forecasts have predictive power for GDP growth forecasts • Benchmark model: AR(4) model of GDP • But: if 2008 crisis is excluded, forecast does not outperform benchmark model any more • In October 2008, I did not need an oil price forecast to tell me we were heading into recession

  13. Why Not Use a Model? (2) • The more general point is: benchmark models pretend that we know less than we actually do • The fact that a candidate variable outperforms the benchmark is too low a threshold

  14. A Case Study: JAPAN

  15. A Case Study: JAPAN

  16. How far Ahead Can We Forecast? • Current-year Forecasts: (current state of macroeconomic foresight) • Short term: next 0-6 months • Medium term: next 3-12 months • Now-cast: previous 3 months • Year-ahead Forecasts • Long term: next 6-24 months

  17. Terminology and Concepts • Know your data • Time of data releases • weekly/monthly/quarterly • REVISIONS • Seasonally Adj? • Real vs nominal magnitudes • Calculating annual changes • oyavs y/y vs q4/q4 • Smoothing (3mma) • Extrapolating annualized quarterly rates from monthly data Real GDP Releases Percent change, q/q, saar

  18. Then How Do We Make Forecasts? Forecasting with Spreadsheets • GDP = Private Consumption • + Government Consumption • + Fixed Investment (Business + Residential + Government) • + Net Exports • + Change in Inventories

  19. Then How Do We Make Forecasts? • 2. Information Gathering • Track Your Data • (Framework to incorporate latest releases quickly)

  20. Then How Do We Make Forecasts? • 3. Incorporate the latest information into the Forecast • benchmark models • judgmental forecast

  21. Short-term • Forecasts based on highest available frequency • Consumption: • Real private consumption index • auto sales, retail sales, store sales • Exports and Imports: • Real export and import indices: • 10-day customs trade • 20-day customs trade • Industrial Production • Business Investment: • IP Report, BoJ Tankan Survey, Cooperation Survey

  22. Medium Term • Leading indicators • Shoko Chukin Survey • Economy watchers Survey • PMI (manufacturing/services) • Consumer Confidence Ind. • Survey of Unemployment Trends • Lending Survey • Political Economy

  23. Medium Term • Watch for base effects (inflation) • mean reversion (inventories like an error term) • statistical carryover

  24. Long-term • Adopt a view • Benchmark: assumptions for potential growth (aging population, declining saving rate) • Adopt a view • Make adjustments for other known developments • Example: • External Factors: Euro Area crisis, China • Domestic Factors: political stability (elections)

  25. How far Ahead Can We Forecast? Japan’s Real GDP Growth Rate Percent change, oya, CY 9.7% 1956-69 5.2% 1970s 3.7% 1980s 1.5% 1990s 0.8% 2000s

  26. A Case Study: JAPAN

  27. Real-Time Forecasts for the World Economy:A Practitioner's Perspective Robin Koepke and Emre Tiftik Global Macroeconomic Analysis Department Institute of International Finance George Washington University Washington, DC April 26, 2012

More Related