Transcript Estimating Supply, Demand, Import and Export

Estimating Import Demand
for Fresh Citrus
Gary D. Thompson
Almuhanad Melhim
The University of Arizona
Linda Calvin
Economic Research Service
Why Study Import Demand?
Quantify Impacts of SPS Measures:
 Elasticities (Own- & Cross-Price)

Flexibility Estimates
 Estimate Welfare of Impacts SPS
Characteristics of SPS Measures
Clementines from Spain
 Country- & Even Region-Specific
 Date-Specific
 Aggregated Data Not Appropriate
U.S. Mandarin Imports, 2000-03
25000
Spain
20000
15000
10000
5000
S. Africa
Australia
Morocco
0
Jan
Feb Mar Apr May
Source: FATUS
Jun
Jul Aug
Sep
Oct Nov Dec
Supply Side Drives Availability
Clementines from Spain
 Zero Import Quantity  No Import Price
Unobservable, Not Unobserved
 Not Censoring; Partial Truncation:
Missing Price & Quantity
 Micro-Data Censoring Models Not Appropriate
Possible Approaches to Truncation
 Incidental Truncation
Sample selection is typically crosssectional.
Sample selection of import availability
depends on agro-climatic factors
(e.g. weather throughout the year)
Truncation & Demand Systems
 Multiple selectivity equations +
Demand system equations
Cross-Sectional - Sequential
Selectivity Models (Lahiri & Song)
Partial Truncation at Product Level
30,000
Mandarins
25,000
20,000
15,000
10,000
5,000
Oranges
Tangerines
0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Selectivity Equations: Probit/Logit
 Binary Regression:
Dep. Vbl. = 0 if no imports
= 1 if positive imports
Exp. Vbls.: Temperature; Precipitation
in Production Region
Marshallian Demand Equations
 Incomplete Demand System
LINQUAD:
Weak integrability guarantees reliable
elasticity and welfare measures.
Demand Equations + Truncation
 Introduce Correction for Partial
Truncation as Demographic Shifter
in LINQUAD:
 Not just inverse Mills ratio
 Multivariate normal is maintained
Choice of Samples for Estimation
1. Consecutive Months Each Year

Oct. - Feb. Season; 1992-93 to 2002-03
2. Aggregate Temporally to Eliminate Zero
Quantities & Missing Prices

Semi-annual; 1989 - 2003
Uncompensated Elasticities, Sample Median
Tangerine
Mandarin
Orange
Tangerine
-0.026
-4.016
0.463
Mandarin
-0.006
-4.375
0.039
Orange
0.374
-0.670
-2.268
Sample: Monthly, Oct.-Feb., 1992 – 2003 (T = 55)
Uncompensated Elasticities, Sample Median
Tangerine
Mandarin
Orange
Tangerine
-2.193
0.543
-0.134
Mandarin
0.020
-0.326
-0.463
Orange
0.133
-0.498
-0.101
Sample: Semi-annual, 1989 – 2003 (T = 30)
Own-Price Elasticity, Mandarins
0
Oct
99
Oct
00
-2.5
-5
-7.5
-10
Suspension of
Spanish Imports
Oct
01
Oct
02
Cross-Price Elasticity, Mand.-Orange
0.25
Suspension of
Spanish Imports
0.2
0.15
0.1
0.05
0
Oct
99
Oct
00
Oct
01
Oct
02
Correction for Truncation
1. Necessary for modeling seasonal
availability of imports (or exports).
2. Yields reasonable, if highly variable,
elasticity estimates.
3. Uses data readily available, e.g. FATUS,
NOAA.
Future Work
 Demand for Domestic & Imported Fresh
Citrus
 Apply to other specialty crops, e.g.
asparagus, fresh tomatoes.
 MLE or Non-Parametric Estimation