Managerial Economics- Demand Forecasting Techniques By Tarun Das

  • Uploaded by: Professor Tarun Das
  • 0
  • 0
  • June 2020
  • PDF

This document was uploaded by user and they confirmed that they have the permission to share it. If you are author or own the copyright of this book, please report to us by using this DMCA report form. Report DMCA


Overview

Download & View Managerial Economics- Demand Forecasting Techniques By Tarun Das as PDF for free.

More details

  • Words: 2,162
  • Pages: 8
Tarun Das

Page 1

11/26/2009

Demand Projections Prof. Tarun Das, IILM, New Delhi-110003. 1. Purpose and Use of Demand Projections 1.1. Demand projections form the basic foundations of corporate planning. 1.2. These are essential for production, pricing and employment planning. 1.3. Every big corporate house employs statisticians and econometricians for analyzing and forecasting market demand. 1.4. Basic building block of demand analysis is the empirical demand functions. 1.5. However, these are subject to identification and specification problems. 2. Dimensions of Demand Projections 2.1. Macro models (National level) 2.2. Meso (Middle level- States) 2.3. Spatial (over space) 2.4. Regional (over regions) 2.5. Industrial (over industries) 2.6. Sectoral (sectors- rural, urban) 2.7. Micro (at unit/firm levels) 2.8. Inter temporal (over time) 2.9. Intergenerational (over generations) 3. Factors Influencing Demand Projections 3.1. Overall objective and purpose- a part of corporate plan 3.2. Planning horizon 3.3. Product disaggregation 3.4. Regional disaggregation 3.5. Quantifiable variables influencing demand and their perspectives in the short and medium term 3.6. Constraints in the system 3.7. Expected change in environment 4.

Forecasting techniques 4.1 Opinion polls- collecting opinions of those who are knowledgeable about the markets such as sales representatives, sales executives, professional marketing experts and consultants. The opinion poll methods include: (a) Expert opinions- although the method is simple and less expensive, it has limitations due to subjective judgments and may lead to either over-estimation or under-estimation. (b) Delhi methodSimilar to expert opinion, but here experts are provided with alternative forecasts and asked to give their expert opinions and revisions, with explanations, if any. These unstructured opinions of experts can be used to cross check results obtained from more sophisticated and statistical techniques. (c) Market surveys- widely used to estimate and forecast demand.

1

Tarun Das 4.2

4.3 4.4

Page 2

Barometer method and Economic indicators- Barometer method is generally used for forecasting weather. The essence of the technique can be used to forecast demand. The basic approach is to use a set of economic indicators to project demand. These indicators can be classified as the following: (a) Leading indicators- those indicators which move up or down ahead of some other series such as net investments, new constructions and transport links, prices of materials, contracts and orders, change of inventories, net corporate profits etc. (b) Coincidental indicators- those which move up or down simultaneously with the level of economic activity such as employees and payrolls, rate of unemployment, gross national product, sales by manufacturing, trading and retail sectors, personal incomes etc. (c) Lagging indicators- those which follow after some time lag such as wage rates and inflation, consumer credits, lending rates, outstanding loans etc. Projections techniques Econometric models

5. Methods of demand Projection 5.1. Average growth approach 5.2. Demand Intensity approach 5.3. Per capita demand approach 5.4. Elasticity approach 5.5. End-Use/ Input-output approach 5.6. Material balance approach 5.7. Time Trend Method 5.8. Engel curves 5.9. Econometric approach- multiple regression approach 5.0 An Example- Trends of Sales, Income and Population Year 2000 2001 2002 2003 2004 2005 Total Average

11/26/2009

Time 1 2 3 4 5 6 6

Sales (S) (Rs.bln) 1200 1300 1400 1550 1680 1800 8930 1488.3

Income (Y) (Rs.bln) 21043 22960 24510 26900 29514 32465 157392 26232

Population (P) (mln) 1019 1037 1055 1072 1088 1104 6375 1062.5

2

Tarun Das

Page 3

11/26/2009

Fig-5.1: Trends of Sales, Income and Population 4000 3000 2000 1000 0 1

2

S (Rs.bln)

3

4

5

Y (Rs.10bln)

6 P (mln)

5.1 Average growth method: Year

S (Rs.bln)

Y (Rs.bln)

P (mln)

GR (S)

2000

1200

21043

1019

8.5

2001

1300

22960

1037

8.3

2002

1400

24510

1055

7.7

2003

1550

26900

1072

10.7

2004

1680

29514

1088

8.4

2005

1800

32465

1104

7.1

Total

8930

157392

6375

50.8

1488.3

26232

1062.5

8.5

Average

Average growth rate of sales= 8.5% Projection of sales for 2006 = 1800 x (1+8.5/100)

= 1800 x 1.085 = 1953 5.2 Demand Intensity Approach Year 2000 2001 2002 2003 2004 2005 Total Average

S (Rs.bln) 1200 1300 1400 1550 1680 1800 8930 1488.3

Y (Rs.bln) 21043 22960

Intensity 0.0570 0.0566

24510 0.0571 26900 0.0576 Ave. Sales Intensity w.r.t. GDP = 0.0567 29514 0.0569 Assumption: GR of GDP in 2006=10% 32465 0.0554 GDP in 2006 = 32465 x 1.10 =35712 157392 0.0567 Proj. for 2006 =35712 x0.0567= 2026 26232 0.0567 5.3 Percapita Demand Approach

Year S (Rs.bln) Y (Rs.bln) P (mln) PC Sales=S/P 2000 1200 21043 1019 1.18 2001 1300 22960 1037 1.25 2002 1400 24510 1055 1.33 2003 1550 26900 1072 1.45 2004 1680 29514 1088 1.54 2005 1800 32465 1104 1.63 Assumption: Population GR in 2006 = 1.5%, Percapita sales = 1.75 Sales Proj. for 2006 = 2005 Pop x Pop Growth factor x PC sales = 1104 x 1.015 x 1.75 = 1961

3

Tarun Das

Page 4

11/26/2009

5.4 Demand elasticity approach Year

S (Rs.bln)

Y (Rs.bln)

P (mln)

GR (S)

GR (Y)

GR (P)

2000

1200

21043

1019

8.5

9.2

1.9

2001

1300

22960

1037

8.3

9.1

1.8

2002

1400

24510

1055

7.7

6.8

1.7

2003

1550

26900

1072

10.7

9.8

1.6

2004

1680

29514

1088

8.4

9.7

1.5

2005

1800

32465

1104

7.1

10.0

1.5

Total

8930

157392

6375

50.8

54.5

10.0

1488.3

26232

1062.5

8.5

9.1

Average

Sales elasticity w.r.t. Y & P = GR (S)/GR (Y) or GR (P)

0.93

1.7 5.0

Sales proj.for 2006: Assume GR (Y)=10%, GR (P)=1.5% (1) Sales GR= 10 x 0.93 = 9.3% (2) Sales GR= 1.5 x 5 = 7.5%

Projection of sales for 2006 = 1967 Projection of sales for 2006 = 1935

5.5 End-Use Method- Planning Commission (PC) Model for Five-Year Plans Demand = Intermediate + Final Demand = Σ aij Xj +Fi, Intermediate Demand = Use by other industries in the process of production = Σ aij Xj, where aij is the amount of ith good used per unit of jth good, Final Demand (Fi) = Private consumption (Ci) + Public consumption (Gi) + Investment (Ii) +Stocks (Sti) +Exports (EXPi) – Imports (IMPi) Fi = Ci + Gi + Ii + STi + EXPi - IMPi 5.6 PC Macroeconomic Model in Leontief Input-Output Framework Di= Xi = Σ aij Xj +Fi Fi = Ci + Gi + Ii + STi + EXPi - IMPi Xi = Supply of ith good Di = demand for ith good Ci estimated by Engel curves Ii is estimated by a distributed lag model on investment. Gi is estimated by minimum needs program and other public distribution and welfare programs of the government. Sti is estimated by fixed coefficients. EXPi are exogenous. IMPi = Σ mj Xj + ki Ci + bi Gi + hi Ii 5.7 Time Trend Method Year 2000 2001 2002 2003 2004 2005

Time (T) 1 2 3 4 5 6

4

S (Rs.bln) 1200 1300 1400 1550 1680 1800

Log (S) 7.09 7.17 7.24 7.35 7.43 7.50

Tarun Das

Page 5

11/26/2009

Fitted Time Trends

Type Linear S = α + β T Exponential Log S = α + β T

α

β

R-SQ

1059.3

122.6

0.996

7.006

0.083

0.998

Projection for 2006 (Time=7)

Linear S = α + β T Exponential Log S = α + β T

1917 1969

5.8 Engel Curves • Engel curves show the relationship between per capita consumption demand and per capita income. • There are various forms of Engel curves. Linear C = α + β Y Loglinear Log C = α + β Log Y Semi log Log C = α + β Y Log Inverse Log C = α + β / Y Log Log Inverse Log C = α + β 1 Log Y+ β 2 / Y • These are estimated on the basis of consumption data obtained from the household expenditure surveys. Year Time S (Rs.bln) Y (Rs.bln) log(s) log(Y) 2000 1 1200 21043 7.09 9.95 2001 2 1300 22960 7.17 10.04 2002 3 1400 24510 7.24 10.11 2003 4 1550 26900 7.35 10.20 2004 5 1680 29514 7.43 10.29 2005 6 1800 32465 7.50 10.39 Assumption: GR of income =10%, Income (Y) in 2006 = 1.10 x 32465 = 35712 Sales Projections for 2006 Linear 1998 Log-linear 2001 Semi log 2075 Log-inverse 1936

Type α β R-SQ

Linear

Log linear

Semi log

77.17

-2.45

6.35

8.27

0.05

0.96

3.6E-05

-24916

0.994

0.995

0.984

0.998

5

Log-inverse

1/ Y 4.75217E-05 4.3554E-05 4.07997E-05 3.71747E-05 3.38822E-05 3.08024E-05

Tarun Das

Page 6

11/26/2009

6.1 Econometric Demand Functions a.Econometric Demand Functions help to establish statistical relations between demand and major factors that influence demand. b. Demand functions can be specified, calibrated, tested, monitored, updated, simulated and predicted with certain degree of confidence. c.Identify potential variables 6.2 Steps in econometric demand analysis i. ii. iii. iv. v. vi. vii.

Have a sound theoretical basis Specify equations Identify equations Calibration of parameters Testing- Various goodness of fit statistics- R-sq, ξ -sq, Theil index etc. Simulation, projections and planning Monitoring, review and updating

Typical Empirical Demand equations D = a + b P + c Y+ d Pr + e N Where D = demand of a good or service P = Price of the good or service Y = Consumer’s income Pr = Price of a related good or service N = A catch-all variable to take care of omitted variables (could be time) b<0 for normal, b>0 for inferior good c>0 for normal, b<0 for Giffen good d>0 for substitute d<0 for complement 6.3 Types of econometric models (a) (b) (c) (d)

Static (at a particular time period) and dynamic (takes care of change of time and business environment) Consistent (consistency among the systems equations), behavioral (on the basis of consumers, producers and traders behavior) and optimizing (maximizing revenue, market share, profits or minimizing costs etc.) Partial equilibrium (deals with specific good) or general equilibrium (considers equilibrium in the whole system) sectoral model (deals with a sector) or the economy-wide model (all sectors)

6.4 Types of data for econometric analysis (a) (b) (c)

Time series (over time- inter-temporal) Cross section (across consumers, regions, countries etc. at a particular time) Pooled- Pooling of consumers and sectors (All consumers in urban, rural)

6

Tarun Das

Page 7

11/26/2009

(d) Panel - Combination of time series and cross section data 6.5 Types of variables

• • • • • • •

Ct = α + β Yt + χ Wt + δ Ct-1 + ε Dt+λ T Dt = 1 if t is 1991 to 2003, 0 otherwise Endogenous variables (determined within the model) Ct, Yt Exogenous/ predetermined variable Wt Parameters α , β , χ , δ , ε , λ Lagged variable Ct-1 Instrumental variable Wt Dummy (binary, categorical, indicative, qualitative, dichotomous) variable Dt Catch all variable (Time T to take care of all omitted variables) 7. Concluding Observations- General Agreements by Modelers a. State your biases, intuitive arguments and limitations of the model. b. Equations provide guideposts and are not expected to produce precise results. c. Econometric results should be supplemented by qualitative judgments for useful policy formulation and corporate planning. d. Social environment and political economy may be treated as given in the model. e. Models should be tested rigorously for the real world with full range of policies. f. Sufficient resources should be used for full documentation of the model, so that any other group can test, calibrate and run the model. g. Documentation should be clear and free from jargons for general understanding by the non-technical audience. h. Modelers should specify data sources and share their data. i. It is needed to continually review, monitor, update, upgrade, and simulate the model. j. However, limited purpose model is better than general-purpose model. k. Descriptive model is better than normative and subjective model. l. Policy oriented model is better than general understanding model. m. Short run and medium term models are better than long-term models. n. Sectoral model is better than economy wide model. o. To guess crucial missing data rather than leaving out. p. Try to deal with future economic agents, technology change, prices and population.

7

Tarun Das

Page 8

11/26/2009

Review Questions 1.

Why is demand forecasting essential for corporate planning? Discuss critically the different opinion poll methods for demand forecasting.

2.

Define leading indicators, coincidental indicators and lagging indicators for demand projections. What would be the appropriate leading indicators, coincidental indicators and lagging indicators for forecasting demands for steel?

3.

What is meant by an Engel curve? Distinguish between time series, cross section, pooled and panel data and their uses for estimation of Engel curves.

4.

You are given the following data: Year

Sales of Cars GDP Population (Million) (Rs. billion) (Million) 2004 10 100 1000 2005 11.5 110 1016 Further assume that GDP is expected to grow at the rare of 10% and population at 1.5% in 2006. Forecast demand for cars in 2006 under the following methods: (a) Elasticity of demand with respect to GDP, (b) Average demand intensity in GDP, (c) Average per capita demand for cars and (d) Past growth rate of cars.

8

Related Documents


More Documents from "Gaurav Kumar"