WatColl Schoengold

Price Elasticity Reconsidered: Panel Estimation of an Agricultural Water Demand Function: 

Price Elasticity Reconsidered: Panel Estimation of an Agricultural Water Demand Function Dr. Karina Schoengold School of Natural Resources Agricultural Economics University of Nebraska, Lincoln

Do Farmers Respond to an Increase in the Price of Water (a necessary input)?: 

Do Farmers Respond to an Increase in the Price of Water (a necessary input)?

Motivation: 

Motivation Agriculture is the largest user of fresh water in many places worldwide Few empirical studies of agricultural water demand exist (Moore, Gollehon and Carey (1994); Ogg and Gollehon (1989)) Economic incentives are increasingly important in resource policy decisions

How Do Farmers Respond to an Increase in the Price of Water?: 

How Do Farmers Respond to an Increase in the Price of Water? Reallocate land (including fallowing) Invest in precision (water-conserving) irrigation systems (drip, sprinkler, central-pivot) Improve management of water resources through better timing of water applications No response

Contributions of This Research: 

Contributions of This Research Estimate direct and indirect price elasticities of water demand using micro-level panel data Correct for the endogeneity of land allocation in water demand estimation

Data: 

Data 8-year balanced panel of crop and irrigation acreage totals for 117 sections (AEWSD) Water use totals by section (AEWSD) Water prices by year (AEWSD) Output prices by year (Kern County) Long-run land quality characteristics Yearly average temperature (WRCC) AEWSD = Arvin Edison Water & Storage District WRCC = Western Regional Climate Center Section = 640 acres of land (not fully in agriculture)

Where is Arvin Edison (AEWSD)?: 

Where is Arvin Edison (AEWSD)?

Data Complications: 

Data Complications Endogenous land allocation (in addition to water use, land use choices are affected by water price) Crop and irrigation technology are jointly determined Lots of corner solutions in acreage totals

Water Rates by Year (in dollars): 

Water Rates by Year (in dollars)

Empirical Strategy: 

Empirical Strategy Step I: Estimation of land allocation using a share model with modifications (Tobit estimation) Step II: Use of predicted values to estimate water demand at the section level Step III: Use results to calculate direct (from step II) and indirect (from step I) water demand elasticities Step IV: Calculate predicted water use with different irrigation technologies (conservation potential)

Number of Sections Choosing a Crop & Irrigation Combination (2001): 

Number of Sections Choosing a Crop & Irrigation Combination (2001)

Step I – Estimation of Land Allocation: 

Step I – Estimation of Land Allocation Using a Tobit Estimation (maximum likelihood), I estimate J equations of the following form: We make the typical assumptions of a Tobit:

Slide13: 

Empirical Results I: Land Allocation (Dependent Variables are the Total Hectares in Each Land Allocation Choice) Results show the estimated coefficient and standard error. Other variables included are minimum wage, all output prices, all lagged acreage values, and a constant.

Slide14: 

Empirical Results II: Land Allocation (Dependent Variables are the Total Hectares in Each Land Allocation Choice) Results show the estimated coefficient and standard error. Other variables included are minimum wage, all output prices, all lagged acreage values, and a constant.

Land Allocation Estimation: Results and Implications: 

Land Allocation Estimation: Results and Implications Land use choices are conditional on observed input and output prices, land quality, and previous investment Different existing conditions (land quality or previous investment) lead to heterogeneous responses to an exogenous change in relative prices

Step II – Estimation of Water Demand: 

Step II – Estimation of Water Demand Variables included are: land quality characteristics, input prices for labor and water, predicted acreage totals, and yearly climate variables I estimate: Defining:

Empirical Results: Water Demand: 

Empirical Results: Water Demand IV results are clustered by section and standard errors are bootstrapped, 936 observations

Step III – Calculation of Direct and Indirect Water Demand Elasticities: 

Step III – Calculation of Direct and Indirect Water Demand Elasticities Indirect effects are determined using the marginal effects from the Tobit estimations, which are calculated using the above expression.

Empirical Results: Elasticity Calculations: 

Empirical Results: Elasticity Calculations Bootstrapped standard errors are in parentheses.

Conclusions and Implications: 

Conclusions and Implications Direct and indirect elasticities of water demand are both negative (indirect effect is about 47% of the total) Validates the use of price as a mechanism to promote conservation Conservation expectations must be conditioned on land quality and existing allocation