Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.
ESTIMATION OF PRODUCTION AND COST FUNCTION <ul><li>For practical decision-making purposes it is necessary to obtain estima...
Estimation of Production and Costs involves: <ul><li>1. Data collection (time series, cross-sectional data). </li></ul><ul...
However, data collection can be difficult  (even more so than in demand estimation) : <ul><ul><li>it may be relatively eas...
Estimation of Production Functions <ul><li>A short review of theory </li></ul><ul><li>SHORT-RUN:  Linear Production Functi...
Estimation of Production Functions  continued <ul><li>Quadratic Production Function </li></ul><ul><li>Q = a + bL - cL 2 </...
Quadratic Production Function   continued Q L L AP MP MP AP a) b)
<ul><li>Cubic Production Function </li></ul><ul><li>Q = a + bL + cL 2  - dL 3 </li></ul><ul><li>MP L = b + 2cL - 3dL 2 </l...
Cubic Production Function   continued Q L L AP MP MP AP a) b)
Production Function as Power Function <ul><li>Q = aL b </li></ul><ul><ul><li>if </li></ul></ul><ul><ul><li>B > 1, Q increa...
Power Function  continued Q L b<1 b=1 b>1
Power Function  continued <ul><li>Power function is the most frequently used type of production function in empirical work...
The Cobb-Douglas Production Function <ul><li>A special case of power functions: </li></ul><ul><li>Q = aL b K 1-b , </li></...
Reformulation by Cobb and Douglas: <ul><li>Q = aL b K c </li></ul><ul><li>b + c = 1, constant returns </li></ul><ul><li>b ...
Properties of  the Cobb-Douglas function that have kept it so popular for 90 years <ul><li>1. Both inputs have to be used ...
<ul><li>3. Allows to investigate MP for any factor while holding all others constant.  So it is useful both in short-run a...
Elasticity of Production <ul><li>Measures the sensitivity of total product to a change in an input in percentage terms: </...
Estimation of Production Functions <ul><li>Regression analysis often used </li></ul><ul><ul><li>time-series or cross-secti...
Estimation of Production Functions  continued <ul><li>Time-series </li></ul><ul><li>1. If data in monetary terms, an infla...
Estimation of Production Functions  continued <ul><li>Cross-sectional </li></ul><ul><li>1. No technological change over ti...
Cost Estimation <ul><li>Short-run cost functions are estimated to help managers to determine optimal pricing policy for th...
<ul><li>Long-run cost functions are used in planning firm’s investment decisions </li></ul><ul><ul><li>To determine the ex...
Estimation of Short-Run Cost Functions <ul><li>Techniques used: </li></ul><ul><li>mostly regression analysis with time ser...
Shapes of Short-Run Cost Functions <ul><li>Cubic cost function: </li></ul><ul><li>A cubic cost function represents the nor...
Cubic cost function  continued <ul><li>TC = a + bQ - cQ 2  + dQ 3 </li></ul><ul><li>AC = a/Q + b - cQ + dQ 2 </li></ul><ul...
Cubic cost function $ TC Q
First marginal and average costs decrease and then increase: $ MC Q AC
Quadratic cost function <ul><li>If data does not fit to a cubic cost function, we can try to fit it to a quadratic one. </...
Costs increasing at increasing rate $ TC Q
No decreasing marginal cost! $ MC Q AC
Linear cost function <ul><li>Also a linear total cost function can be fitted. Then the three functions get the following f...
No Law of Diminishing Marginal Returns $ TC Q $ MC Q AC
Note! <ul><li>Many empirical short-run cost studies have found a linear relationship between total cost and output, indica...
So… <ul><li>Should economist revise their view of U-shaped average and marginal cost curves? </li></ul><ul><ul><li>data em...
Estimation of Long-Run Cost Functions <ul><li>Techniques used: </li></ul><ul><li>Regression analysis </li></ul><ul><li>Eng...
Regression analysis in L-R Cost Estimation <ul><li>Mostly with cross-sectional data </li></ul><ul><li>Pluses: </li></ul><u...
Regression analysis in L-R Cost Estimation  continued <ul><li>Minuses: </li></ul><ul><ul><li>interregional cost difference...
Engineering cost method <ul><li>Based on understanding of inputs and outputs and their relationships </li></ul><ul><li>In ...
Engineering cost method  continued <ul><li>Pluses: </li></ul><ul><ul><li>technology held constant </li></ul></ul><ul><ul><...
Survivor technique <ul><li>This method, suggested by G. Stigler, bases its findings on the change in the proportion of tot...
Survivor technique  continued <ul><li>Pluses: </li></ul><ul><ul><li>simple </li></ul></ul><ul><ul><li>avoids unreliable da...
Estimation Of Production And Cost Function
Upcoming SlideShare
Loading in …5
×

Estimation Of Production And Cost Function

45,151 views

Published on

Estimation Of Production And Cost Function

  1. 1. ESTIMATION OF PRODUCTION AND COST FUNCTION <ul><li>For practical decision-making purposes it is necessary to obtain estimates of production and cost functions. </li></ul><ul><li>In economics, it is usually hard to perform controlled laboratory experiments. Instead, actual operating data are used with some statistical procedures to derive these estimates. </li></ul>
  2. 2. Estimation of Production and Costs involves: <ul><li>1. Data collection (time series, cross-sectional data). </li></ul><ul><li>2. Have to assume some mathematical form for the function. </li></ul><ul><li>3. Have to determine the estimation method for finding the parameter values (regression analysis for instance). </li></ul>
  3. 3. However, data collection can be difficult (even more so than in demand estimation) : <ul><ul><li>it may be relatively easy to estimate the use of labor in production, but to estimate capital usage can be very difficult </li></ul></ul><ul><ul><li>most cost data are obtained from accounting records, so they do not necessarily conform to costs as defined by economists </li></ul></ul><ul><ul><li>how to treat the assumption of constant technology for a given production or cost function </li></ul></ul>
  4. 4. Estimation of Production Functions <ul><li>A short review of theory </li></ul><ul><li>SHORT-RUN: Linear Production Function </li></ul><ul><li>Q = a + bL </li></ul><ul><li>- very simple, but does not take into account the law of diminishing returns </li></ul>L Q = a + bL MP L Q
  5. 5. Estimation of Production Functions continued <ul><li>Quadratic Production Function </li></ul><ul><li>Q = a + bL - cL 2 </li></ul><ul><li>MP L = b - 2cL </li></ul><ul><li>AP L = a/L + b -cL </li></ul><ul><li>- implies diminishing returns, but not rising marginal product at the beginning </li></ul>
  6. 6. Quadratic Production Function continued Q L L AP MP MP AP a) b)
  7. 7. <ul><li>Cubic Production Function </li></ul><ul><li>Q = a + bL + cL 2 - dL 3 </li></ul><ul><li>MP L = b + 2cL - 3dL 2 </li></ul><ul><li>AP L = a/L +b + cL - dL 2 </li></ul><ul><li>- implies first increasing marginal returns and then diminishing returns </li></ul>
  8. 8. Cubic Production Function continued Q L L AP MP MP AP a) b)
  9. 9. Production Function as Power Function <ul><li>Q = aL b </li></ul><ul><ul><li>if </li></ul></ul><ul><ul><li>B > 1, Q increasing at increasing rate: MP L increasing </li></ul></ul><ul><ul><li>B = 1, Q increasing at constant rate: MP L constant </li></ul></ul><ul><ul><li>B < 1, Q increasing at decreasing rate: MP L decreasing </li></ul></ul><ul><ul><li>Major advantage of the power function is the fact that it can be transformed in a log-linear function </li></ul></ul><ul><ul><li>log Q = log a + b log L </li></ul></ul>
  10. 10. Power Function continued Q L b<1 b=1 b>1
  11. 11. Power Function continued <ul><li>Power function is the most frequently used type of production function in empirical work, even though it cannot exhibit two directions for marginal product on the same function. </li></ul><ul><li>One reason for its popularity is that it can be readily transformed into a function with two or more independent variables: </li></ul>
  12. 12. The Cobb-Douglas Production Function <ul><li>A special case of power functions: </li></ul><ul><li>Q = aL b K 1-b , </li></ul><ul><li>Original version with constant returns to scale ( b + 1 - b = 1) introduced by Cobb in 1928 </li></ul><ul><li>He estimated the production function of U.S. manufacturing output for years 1899-1922 </li></ul>
  13. 13. Reformulation by Cobb and Douglas: <ul><li>Q = aL b K c </li></ul><ul><li>b + c = 1, constant returns </li></ul><ul><li>b + c > 1, increasing returns </li></ul><ul><li>b + c < 1, decreasing returns </li></ul><ul><li>Can only use one of these at a time… so which one to choose? </li></ul>
  14. 14. Properties of the Cobb-Douglas function that have kept it so popular for 90 years <ul><li>1. Both inputs have to be used simultaneously to get an output </li></ul><ul><li>2. Can exhibit different returns to scale (even though can not show a unit or an industry to move through all three stages) </li></ul>
  15. 15. <ul><li>3. Allows to investigate MP for any factor while holding all others constant. So it is useful both in short-run and long-run analysis. </li></ul><ul><li>4. Elasticities are equal to the exponents b and c. (constant in this formulation) </li></ul>
  16. 16. Elasticity of Production <ul><li>Measures the sensitivity of total product to a change in an input in percentage terms: </li></ul>
  17. 17. Estimation of Production Functions <ul><li>Regression analysis often used </li></ul><ul><ul><li>time-series or cross-sectional analysis? </li></ul></ul><ul><li>Both methods have their advantages and disadvantages </li></ul>
  18. 18. Estimation of Production Functions continued <ul><li>Time-series </li></ul><ul><li>1. If data in monetary terms, an inflation adjustment is necessary. </li></ul><ul><li>2. Technology may change over time. </li></ul><ul><li>3. Production function assumes that production takes place where input combination is most efficient. </li></ul>
  19. 19. Estimation of Production Functions continued <ul><li>Cross-sectional </li></ul><ul><li>1. No technological change over time, but all plants in the investigation are assumed to have same technology. </li></ul><ul><li>2. Adjustments across different geographical areas must be made. </li></ul><ul><ul><li>wages and price level </li></ul></ul><ul><li>3. No guarantee that each plant operates at the most efficient input combination for the period examined. </li></ul>
  20. 20. Cost Estimation <ul><li>Short-run cost functions are estimated to help managers to determine optimal pricing policy for the company </li></ul><ul><ul><li>used to determine marginal cost of producing additional units of output </li></ul></ul>
  21. 21. <ul><li>Long-run cost functions are used in planning firm’s investment decisions </li></ul><ul><ul><li>To determine the extent of economies and diseconomies of scale in order to select the optimal plant size </li></ul></ul>
  22. 22. Estimation of Short-Run Cost Functions <ul><li>Techniques used: </li></ul><ul><li>mostly regression analysis with time series data </li></ul><ul><li>problems and adjustments </li></ul><ul><ul><li>economic vs. accounting costs </li></ul></ul><ul><ul><li>rate changes; such as tax rates, social security contributions etc. </li></ul></ul><ul><ul><li>output homogeneity </li></ul></ul><ul><ul><li>timing of costs </li></ul></ul><ul><ul><li>accounting changes (have deprecation methods changed…) </li></ul></ul>
  23. 23. Shapes of Short-Run Cost Functions <ul><li>Cubic cost function: </li></ul><ul><li>A cubic cost function represents the normal theoretical cost function, which exhibits both decreasing marginal and average costs and increasing marginal and average costs </li></ul>
  24. 24. Cubic cost function continued <ul><li>TC = a + bQ - cQ 2 + dQ 3 </li></ul><ul><li>AC = a/Q + b - cQ + dQ 2 </li></ul><ul><li>MC = b - 2cQ + 3dQ 2 </li></ul>
  25. 25. Cubic cost function $ TC Q
  26. 26. First marginal and average costs decrease and then increase: $ MC Q AC
  27. 27. Quadratic cost function <ul><li>If data does not fit to a cubic cost function, we can try to fit it to a quadratic one. </li></ul><ul><li>Quadratic cost function: </li></ul><ul><li>TC = a + bQ + cQ 2 </li></ul><ul><li>AC = a/Q + b + cQ </li></ul><ul><li>MC = b + 2cQ </li></ul>
  28. 28. Costs increasing at increasing rate $ TC Q
  29. 29. No decreasing marginal cost! $ MC Q AC
  30. 30. Linear cost function <ul><li>Also a linear total cost function can be fitted. Then the three functions get the following form </li></ul><ul><li>TC = a + bQ </li></ul><ul><li>AC = a/Q + b </li></ul><ul><li>MC = b </li></ul>
  31. 31. No Law of Diminishing Marginal Returns $ TC Q $ MC Q AC
  32. 32. Note! <ul><li>Many empirical short-run cost studies have found a linear relationship between total cost and output, indicating a constant marginal cost. </li></ul>
  33. 33. So… <ul><li>Should economist revise their view of U-shaped average and marginal cost curves? </li></ul><ul><ul><li>data employed concentrate on output levels of limited range </li></ul></ul><ul><ul><li>capital inputs may not be fixed even in short-run </li></ul></ul><ul><ul><li>regression is not a perfect tool </li></ul></ul>
  34. 34. Estimation of Long-Run Cost Functions <ul><li>Techniques used: </li></ul><ul><li>Regression analysis </li></ul><ul><li>Engineering cost method </li></ul><ul><li>Survivor technique </li></ul>
  35. 35. Regression analysis in L-R Cost Estimation <ul><li>Mostly with cross-sectional data </li></ul><ul><li>Pluses: </li></ul><ul><ul><li>since data comes from different firms, quantity of output can vary over relatively wide ranges </li></ul></ul><ul><ul><li>all data from same point of time, so technology will not change </li></ul></ul><ul><ul><li>do not have to regard price changes </li></ul></ul>
  36. 36. Regression analysis in L-R Cost Estimation continued <ul><li>Minuses: </li></ul><ul><ul><li>interregional cost differences </li></ul></ul><ul><ul><li>all firms not necessarily operating at optimal level of technology </li></ul></ul><ul><ul><li>costs may be recorded differently in different firms </li></ul></ul><ul><ul><li>different companies may pay their cost factors differently </li></ul></ul>
  37. 37. Engineering cost method <ul><li>Based on understanding of inputs and outputs and their relationships </li></ul><ul><li>In this approach, the analysis begins with an ”engineering production function”: optimal production input combinations for producing any given level of production is identified. </li></ul><ul><li>Cost can be obtained by multiplying each level of input usage by current price of the input and summing over the inputs. </li></ul>
  38. 38. Engineering cost method continued <ul><li>Pluses: </li></ul><ul><ul><li>technology held constant </li></ul></ul><ul><ul><li>no problem with inflation (current input prices) </li></ul></ul><ul><ul><li>less error from measurement </li></ul></ul><ul><li>Minuses: </li></ul><ul><ul><li>cost estimates are normative </li></ul></ul><ul><ul><li>only direct output costs are estimated </li></ul></ul><ul><ul><li>often made based on pilot plant operations, not actual production </li></ul></ul>
  39. 39. Survivor technique <ul><li>This method, suggested by G. Stigler, bases its findings on the change in the proportion of total industry output produced by firms of different size categories </li></ul><ul><ul><li>look at company size that is successful in an industry! </li></ul></ul><ul><li>Used for deciding optimal plant size </li></ul>
  40. 40. Survivor technique continued <ul><li>Pluses: </li></ul><ul><ul><li>simple </li></ul></ul><ul><ul><li>avoids unreliable data </li></ul></ul><ul><li>Minuses: </li></ul><ul><ul><li>no help in measuring costs for planning purposes </li></ul></ul><ul><ul><ul><li>just tells which company size appears to be more efficient </li></ul></ul></ul><ul><ul><li>implicitly assumes that the industry highly competitive, so survival and prosperity are solely a function of efficient use of resources, not the market power or erection of barriers of entry </li></ul></ul>

×