Sie sind auf Seite 1von 18

Introduction to Forecasting

Outline: 1. What is forecasting?

2. Why use forecasting techniques?


3. Applications of forecasting methods 4. Types of forecasting models 5. Time series vs. cross-sectional data.

Initial points:

The economics profession is held in low regard in some


quarters, partly because there is the widespread perception (if not reality) that economists fail in their most meaningful or potentially useful task, which is to produce accurate forecasts of GDP, inflation, housing starts, stock market movements, interest rates, employment, car sales, imports, . . . .

Econometric forecasting has been, to some extent,


oversold.

Which way is the stock market headed, Professor? If I could see the future, I would be on the plane to Las Vegas

S & P 500, Index of Total Return


4600

Actual Predicted

4400

4200

4000

3800 99:07 99:09 99:11 00:01 00:03 00:05 00:07

Why use forecasting?


Virtually every significant business decision (e.g., decisions to place orders for raw materials, semi-finished articles, or finished goods, borrowing decisions, hiring decisions, decisions to close plants, decisions to open new stores or production facilities, decisions to purchase capital goods, or decisions about near-term production flows) are necessarily based on views of decision-makers about the future course of variables which affect their business. Thus, decision-makers are compelled to form expectations--indeed, to make a forecast-- about future business conditions. Econometric forecasting, for all its shortcomings, beats the alternatives.

A review of forecasting applications


Operations planning and control. Firms use forecasting for inventory and production management, and sales force management,. Strategic decision making. Forecasting is helpful to decision-making with respect to the development of new products, entry into new geographic markets, and so on. Cost estimation: In making bids for construction projects or the delivery of capital goods such as turbine generators, aircraft, or ships, firms must forecast the prices of raw materials , semi-finished articles, and so forth.

Marketing. Decision-makers rely heavily on forecasts about the responsiveness of consumers or buyers to various marketing schemes such as couponing, rebates, tie-ins, low APR financing, volume discounts, and so on. Fiscal and Monetary Policy. Federal Reserve officials, Congress, and the President rely on forecasts of output, employment, inflation, and other macroeconomic variables to develop monetary policy initiatives, tax policy, and government spending policy.
Financial Market Speculation: Money managers (professional and otherwise) rely on forecasts of firms earnings and movements of composite indices (such as the DOW, NASDAQ, or Russell 2000) in formulating their portfolio strategies.

Investment, capacity planning. Decisions to add or subtract capacity depend on long-term forecasts of market conditions. What are the trends in market size and market share?

Business and Government Budgeting. Government appropriations are conditioned upon estimates of revenues from income taxes, excise taxes, etc. The Department of Treasury uses forecasting techniques to assess if a particular tax initiative is revenue neutral, revenue enhancing, or revenue decreasing.
Demography: Demographers use forecasting techniques to forecast population growth around the world, changes in the age composition of the population, and changes in other demographic characteristics of the population.

S&K note that many professional forecasters are momentum followers--i.e., they will tend to extrapolate recent conditions into the near future--and hence miss significant changes in conditions--this habit produces substantial forecast errors. Professional forecasters would rather fail conventionally than succeed unconventionally. This factor produces a tendency toward a convergence of forecasts among professional forecasters.

S&K note that comparatively small forecast errors (that is , from point estimates) can have significant negative consequences for decision-makers. Example: Suppose that the growth of demand for electricity for a power company was forecast to be 6 percent per year but turned out to be 4 percent per year. Building capacity to meet the forecast would generate excess capacity and the error would be compounded over time.

Forecast Error Compounded Over Time Forecasted 1 Demand (Average Megawatt Hours) 10,000 10,600 11,236 11,910 12,625 13,383 14,186 10,400 10,816 11,248 11,698 12,166 12,653 400 420 662 927 1,217 1,533 Actual Demand 2 (Average Megawatt Hours) Average Excess Capacity (Megawatt Hours)

Year 1 2 3 4 5 6 7
1

6 percent per year 2 4 percent per year

Types of forecasting models


Sherman & Kolk (S&K) distinguish between two types of models:
Qualitative models These models are constructed using surveys of experts in specific areas such as residential construction , petroleum/natural gas, aluminum, banking, or the auto industry. Quantitative models: There are two types of quantitative modelstime series models, which are based exclusively on time series data; and causal models. Causal models are based on the analysis of time series data or cross-sectional data.

Time -series data: historical data--i.e., the data sample consists of a series of daily, monthly, quarterly, or annual data for variables such as prices, income , employment , output , car sales, stock market indices, exchange rates, and so on.

Cross-sectional data: All observations in the sample are taken from the same point in time and represent different individual entities (such as households, houses, etc.)

Time series data: Daily observations, Korean Won per dollar


Year Month Day Won per Dollar 1997 3 10 877 1997 3 11 880.5 1997 3 12 879.5 1997 3 13 880.5 1997 3 14 881.5 1997 3 17 882 1997 3 18 885 1997 3 19 887 1997 3 20 886.5 1997 3 21 887 1997 3 24 890 1997 3 25 891

Example of cross sectional data Student ID 777672431 231098765 111000111 898069845 Sex M M F F Age 21 28 19 22 Height 61 511 58 54 Weight 178 lbs. 205 lbs. 121 lbs. 98 lbs.

000341234

20

62

183 lbs

With time series models, we seek to uncover trends in past data and then Time series models include: project them into the future Nave extrapolation Simple moving average models

Complex moving average models (such as ARMA and ARIMA)


The multiplicative time series model Exponential smoothing

Y =f(X)
This class of models entails the estimation of causal relationships (as suggested by theory, usually) between a dependent variable and one or more independent or explanatory variables.The principal tool of causal forecasting is regression analysis.

Example: The Demand for Coal

To forecast the demand for coal, we insert forecasted values of FIS, FEU, etc. into this equation

COAL = 12,262 + 92.43FIS + 118.57FEU 48.90PCOAL + 118.91PGAS


COAL is monthly demand for bituminous coal (in tons)
FIS is the Federal Reserve Board Index of Iron and Steel production. FEU the FED Index of Utility Production.

PCOAL is a wholesale price index for coal.


PGAS is a wholesale price index for natural gas.

Source: Pyndyck and Rubinfeld (1998), p. 218.

Das könnte Ihnen auch gefallen