9. Valuing Impacts from Observed Behavior: Direct Reading: BGVW, Chapters 13
1. Introduction Need to measure social surplus. These are found as areas under demand and supply curves. Triangles, trapezoids and so forth. Curves are usually unknown, so analyst need to find ways to estimate them the curves and, in turn, calculate the triangles. Well look at some direct estimation techniques and then some indirect ways to measure shadow values -- infer values indirectly from other markets and behaviors.
2. Direct Estimation One point plus slope or elasticity. Usually know one point (current price & quantity). If slope is also known (literature on demand for the good you are interested in studying), then it is a simple matter to extrapolate the demand curve. in the basic demand equation below, for example, may be estimated from various studies price quantity slope
If so, it is a simple matter to sketch out the shape of the curve around this point. price p0 quantity q0
If so, it is a simple matter to sketch out the shape of the curve around this point. price p0 quantity q0
Or, may just use the demand equation Or, may just use the demand equation. Plug in current p and q to get intercept . Then you have an estimate of the equation.
Refuse fee example from book Refuse fee example from book. Current price is zero, town plans to raise fee to $.05 per pound. Graph: Social cost is $.06 initial zero price point Robin Jenkins. The Economics of Solid Waste Reduction: The Impact of User Fees. 1993 -- -.40 lb/p/d Note: Validity.
Elasticities and linear demand Elasticities and linear demand. If you know demand is linear and you have an elasticity measure instead of a slope, you can calculate the slope as You need to know price at which elasticity was calculated in the original study. Same basic strategy can be followed for constant elasticity demand curves Discuss example in book (p. 324)
Extrapolating from a few points. You know a few points of p,q from historic data and you try to “fit ” a line through these points. Usually 2 or 3 points. Example from book Known data points
Notes sensitivity to functional form controlling for other factors validity of measures from only 2 points Extrapolating beyond ‘relevant range’
Many Points/Econometrics Suppose you have many observations of price and quantity from a time series or cross sectional data set. price quantity
Regress q on p, or fit a line to this scatter of points Regress q on p, or fit a line to this scatter of points. You’d have something like the red linear regression or blue non-linear regression price quantity
Use the estimated line in your analysis Types of data individual vs. aggregate time series (annual/monthly) vs. cross sectional examples: electricity, farm crops, water, consumer goods galore, automobiles, oil, minerals markets, traffic, computers, wine …… Major econometric issues (1) Omitted variable bias (2) Autocorrelation in time series data
(3) Identification Problem: Are you estmating a deamnd curve or a supply curve when you estimate the relationship between p and q? The ideal case. (Rain example. Cross sectional stories.) Shifting supply is sketching out a demand curve. Stable demand curve Shifting supply curves price quantity
(3) Identification Problem: Are you estmating a deamnd curve or a supply curve when you estimate the relationship between p and q? The ideal case. (Rain example. Cross sectional stories.) Shifting supply is sketching out a demand curve. Stable demand curve Shifting supply curve price quantity Observations
But it could have been supply that was sketched out price quantity
Simultaneous equations/instrumental variables Trouble and likely outcome is that you observe price and quantity relationships that get both shifts. Simultaneous equations/instrumental variables price quanity