Presentation is loading. Please wait.

Presentation is loading. Please wait.

World Bank Institute 2 nd Annual GNPBO Conference Case Study: Forecasting Commodity Revenues Sahir Khan Senior Fellow, University of Ottawa Jason Jacques.

Similar presentations


Presentation on theme: "World Bank Institute 2 nd Annual GNPBO Conference Case Study: Forecasting Commodity Revenues Sahir Khan Senior Fellow, University of Ottawa Jason Jacques."— Presentation transcript:

1 World Bank Institute 2 nd Annual GNPBO Conference Case Study: Forecasting Commodity Revenues Sahir Khan Senior Fellow, University of Ottawa Jason Jacques Director, Canadian Parliamentary Budget Office June 11, 2014

2 “All models are wrong, but some are useful” British Statistician George Box 1919 – 2013 British Statistician George Box 1919 – 2013

3 Transparency Matters PBOs play an important role in enhancing the credibility and transparency of government fiscal and economic forecasts

4 Data Sources The PBO has been provided with a broad mandate to support Parliament and parliamentarians in holding the government to account for the good stewardship of public resources

5 Commodity revenues can be volatile. On the other hand, spending needs for hospitals, roads and police are not. Alberta Budget

6 The Problem  How to forecast government revenues from oil and gas extraction? Parliaments are need at least one data point, and are well served by a second. PBO’s are uniquely placed to fulfill this role. T HREE P ARTS 1.What is forecast price for commodity? 2.What is forecast production? 3.What is the government’s share of revenues?

7 Which Forecasting Approach? The best approach depends on the accessible data and expertise that can be brought to bear on the problem Consensus Judgment What do other experts think Easiest, but with least understanding Trend Analysis History predicts the future Only requires historical aggregate data Fundamental Structured model of relationships Needs historical micro-data of variable in model

8 “In God we trust; all others must bring data” U.S. Statistician W. Edwards Deming 1900 – 1993

9 Data Sources : Prices Data SetsSources  National statistical agency (historical)  Commodity exchange (futures markets)  External forecasts (e.g. International Energy Agency; World Bank, private firms) Wide range of general forecasts for commodity prices, but all need to reflect local pricing conditions

10 Data Sources : Prices Forecasting a “price” will require judgment on the part of forecasters to reflect tacit knowledge of local conditions

11 Data Sources : Production Data SetsSources  National statistical agency (historical)  Firm-level production data (historical/planned)  Independent estimates of potentially recoverable resource Output estimates will vary from jurisdiction to jurisdiction and project to project. Micro-data are necessary to prepare an accurate forecast

12 Data Sources : Production The PBO is well placed to offer a risk analysis of potential output, based on previous forecast errors from publicly available sources  Primary reliance on firm-reported data, including projections  Audited financial results

13 Data Sources : Government Revenues Data SetsSources Government share of commodity output value  Public Accounts of Government (historical)  Government’s budget (forecast)  Royalty agreements and contracts (from government or firms) Royalty regimes will vary from jurisdiction to jurisdiction, interpreting local contracts will require specialized expertise

14 Data Sources : Government Revenues Keeping track of previous forecast errors is a good starting point for assessing the reasonableness of government figures

15 Role of the PBO Risk Analysis of the Government Forecast Positive: “data are not available to assess the validity of the forecast/”we believe there is downside risk” Normative : Quantitative calculation of variance based on previous forecast errors/testing “reasonableness” of key assumptions Independent Forecast Often easier than expected, as everyone faces similar challenges Can provide legislators with better understanding of commodity fundamentals PBO is well-placed to validate Government’s estimate and can often provide important context regarding the risk associated with revenue forecasts.

16 Thank-you Sahir Khan Visiting Senior Fellow Jean-Luc Pépin Chair on Canadian Government Jason Jacques Director of Economic and Fiscal Analysis Canadian Parliamentary Budget Office

17


Download ppt "World Bank Institute 2 nd Annual GNPBO Conference Case Study: Forecasting Commodity Revenues Sahir Khan Senior Fellow, University of Ottawa Jason Jacques."

Similar presentations


Ads by Google