Short-term agricultural supply response to
international food prices and price volatility
Mekbib Haile, Matthias Kalkuhl and Joachim von Braun
Centre for Development Research (ZEF)
February 1, 2013
Outline
 Introduction
 State
of the art
 Theoretical framework
 Research method and data
 Findings
 Conclusions
2
Background
Food prices to remain at high level in the near future?
• (Recent) high prices characterized by more frequent
volatilities and spikes
• Effects on global food supply
•
3
Why do high prices and volatility matter?
High prices
-
Pure consumers
High price volatility
±
+
?
Producers
(net buyers)
-
-
Producers
(net sellers)
Agricultural supply
and food security
4
Objective
Impact of international food prices and their
volatility on inter- and intra-seasonal agricultural
supply response
-
How does information in the course of a year shape the
formation of price expectations and affect acreage adjustment?
-
Acreage response effects of intra-annual international food price
volatility?
-
Food security implications of such intra-annaul acreage
adjustments?
5
Seasonality of agriculture matters
70000
Wheat
Corn
Soybeans
(2008)
Rice
• The global crop production
concentrated to few months
Area planted (1000 Ha)
60000
50000
40000
30000
• No major planting &
harvesting for about a third
of the year, Dec to Mar.
20000
10000
0
350000
Production (1000 MT)
300000
• Nor is production evenly
distributed across
geographical regions 
250000
200000
150000
100000
50000
0
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Source: Own calculations based on global crop calendar information, data
from FAO (2012) and national data sources.
Dec
6
Spatially concentrated
100%
Global sown area share of major producer countries (2008)
Russia
Indonesia
Thiland
75%
Bangladesh
Myanmar
Brazil
Mexico
50%
India
China
EU
25%
Argentina
Australia
USA
0%
Wheat
Corn
Soybean
Rice
Close to 90% for soybeans, and 60% for each of wheat and corn of the global
cultivation is found in the top 5 or 6 growing countries/regions. Nearly half of the global
rice is cultivated in China & India.
7
State of the art
Two general frameworks for analyzing agri. supply response
1) The Nerlovian partial adjustment approach (Nerlove, 1956).
−
−
No detailed input data
Incorporation of producers economic behavior is adhoc
Recent applications:
−
−
−
Kanwar & Sadoulet (2008): output response of cash crops in India
Vitale et al. (2009): supply response of staple crops in Southern Mali
Yu et al. (2012): acreage and yield response of different winter and summer
season crops for the province of Henan in China.
8
Cont‘d
2) The supply function approach (Ball, 1988)
−
Based on profit-maximizing framework and hence integrates producer’s
economic behavior in a theoretically consistent mode
Recent applications:
−
−
−


Coyle (2005): Dynamic yield response, for Manitoba in Canada
Lin & Dismukes (2007): investigates the role of risk in farmers’ acreage
decisions for major field crops in the North Central region of the US
Liang et al. (2011): supply response in the Southeastern US agriculture
Previous studies have micro, sub-national or national scope
Also, they are on annual bases and fail to look into possible
intra-annual acreage adjustments
9
Theoretical Model

10
Emprical Model
We accounted for seasonality in several ways
1. Alternative proxies for price expectations
Spot prices during planting months
 Harvest-time futures prices quoted during planting months

2. Price risk measures
a) Crop price volatility

SD of changes in log price (price-returns): we consider the variances of
the price returns in the 12 months until the month when planting starts
b) Co-variances between expected crop prices
11
Cont‘d

12
Data and sources
Comprehensive data for the period 1961-2010






Planted acreage : National statistics
Harvested acreage: FAOSTAT, USDA
Crop Calendar: FAO GIEWS; USDA
Input and output spot prices: World Bank
Futures prices: Bloomberg
Consumer Price Index: US bureau of Labor statistics
13
Results and discussion
Annual model
Intra-annual models
Model I
Model II
Model III
Dep. variable
Annually aggregated
sown acreage
Monthly sown
acreage
Annual sown acreage
for typical planting
months
Proxies for price
expectations
Annual average
Monthly prices at
prices at the previous planting time
harvest period
Monthly prices at
planting time
Price risk
Annualized SD of
price returns in the
previous year
SD of price returns
in the 12 months
preceding the start
of the planting
period
SD of price returns
in the 12 months
preceding the start
of the planting
period
Result
Conventional annual
acreage elasticties
Average monthly
acreage responses
Month-specific
acreage response
Note: All area and price variables are in logarithmic form
14
Model (#1) (Annual, first differenced)
Variables
Wheat
Own acreage (t-1)
-0.252**
Wheat price
0.069**
Corn price
0.004
Soybean price
0.012
Rice price
Own price volatility
0.015
Fertilize price index
-0.028**
N
*p<0.10, ** p<0.05, *** p<0.01
Corn
-0.281**
-0.100***
0.174***
-0.014
Soybeans
-0.381*
0.036
-0.149*
0.244***
-0.985**
0.012
-0.142
-0.037
Rice
-0.22
-0.054*
0.039
0.004
0.027*
-0.283*
0.014
48
►Newey-West autocorrelation adjusted Standard Errors are employed
15
Intra-seasonal model (#2)
Variable
Own acreage (t-12)
Wheat price
Corn price
Soybean price
Rice price
Wheat price vol.
Corn price vol.
Soybean price vol.
Rice price vol.
Time trend
N
Wheat
0.837***
0.068**
0.021
-0.057**
-0.023
-0.894**
1.014**
0.635*
-0.151
0
Corn
0.842***
0.031
0.113**
-0.015
-0.028**
0.269
0.286
-0.184
0.072
0.001
588
Soybeans
0.961***
0.018
-0.085**
0.111***
0.011
0.307
0.057
0.696
-0.786**
0.002**
Rice
0.628***
-0.007
-0.002
0.007
0.016**
-0.321*
0.128
0.297*
0.095
0.003***
587
Notes: Monthly dummies were also included for each crop regression;
*p<0.10, ** p<0.05, *** p<0.01
16
Intra-seasonal model (#3)
Apr
Wheat
May
Apr
Corn
May
Nov
Nov
Own acreage (t-12)
0.506***
0.468***
0.822***
0.334***
0.346***
0.640***
Wheat price
Corn price
0.134***
-0.05
0.303**
-0.144
0.014
-0.018
-0.044
0.111*
-0.015
0.095**
0.047
-0.116*
Soybean price
0.048
0.023
0.028
0.05
0.02
0.039
Own price volatility
-0.627
-1.807
-0.252
-0.675
-0.082
0.937
Fertilizer price
Time trend
-0.023
0.002**
-0.069**
0.005***
-0.004
0.002
-0.006
0.007***
-0.008
0.008***
0.022
0
Intercept
-0.656
N
* p<0.10, ** p<0.05, *** p<0.01
-6.180***
-1.24
-8.470***
-8.661***
3.462**
Variable
49
17
Intra-seasonal model (#3)
Variable
Own Acreage (t-12)
Wheat price
Corn price
Soybean price
Rice price
Own price volatility
May
Soybeans
Jun
Nov
May
0.792***
0.061
-0.156**
0.178***
0.804***
0.025
-0.13
0.222***
0.923***
0.141
-0.068
0.013
0.662***
0.158
Fertilizer price
-0.039
Time trend
0.003
intercept
-4.888
N
* p<0.10, ** p<0.05, *** p<0.01
-0.169
0.659
0.019*
0.113
-0.034
0.006***
-11.005***
-0.005
0.004
-7.741
0
0.001***
0.431
Rice
Jun
Nov
0.680*** 0.556***
0.022**
0.131
-0.01
-0.213
-0.006
0.002***
-0.51
0
0.002
0.464
49
18
Conclusions

Estimation of annual and intra-annual models reveal that global
acreage responds to international prices
−

Comparison of the annual and the monthly acreage response
elasticities suggests that acreage adjusts seasonally around the
globe to new information and expectations.
−

Use of country-specific crop calendar
Given the seasonality of agriculture, time is of an essence for acreage
response
The analysis indicates that acreage allocation is more sensitive to
prices in the northern hemisphere spring than in winter and the
response varies across months.
19
Cont‘d

International food price uncertainty affects aggregate supply response

In summary, own crop price volatility seems to have a negative impact on
wheat, corn and rice acreages but no or little impact on soybean area.

The may imply that the behavioral assumption of risk-aversion is likely to
hold for the majority of wheat, corn and rice producers in the world
•

However, the majority of the global soybean farmers seem not to be unwilling to take
price risks to acquire the associated higher returns of agricultural investments.
This is relevant for policy makers suggesting that reducing output price
volatility leads to an expansion of agricultural land and hence crop
production; however, it may have unexpected and possibly undesirable
outcome for some crop producers.
20
Relevance and outlook

The short-term supply analysis provides a first step for establishing a
global short-term supply model that predicts area planted (and, thus,
expected harvest) according to current world market prices.

Besides indicating potential food supply shortages, such supply
model helps to analyze whether current prices are consistent with
‘fundamentals’ or whether they are driven by speculation or trade
disruption.
We further consider a cross-country dynamic panel model for
acreage and yield


Consideration of the effects of weather condition variables (El Niño
and La Niña)
21
Thank you
Comments and Suggestions!
mekhaile@uni-bonn.de
Acreage response: Panel data
Variable
Wheat
(1)
Corn
(2)
(1)
(2)
Lagged own area
0.856*** 0.895***
0.981*** 0.982***
Wheat spot price
0.099***
-0.043
Wheat futures price
0.112***
0
Corn spot price
-0.001
0.087**
Corn futures price
0.119**
0.053
Soy spot price
-0.019
0
Soy futures price
-0.129**
-0.062
Rice spot price
Wheat price vol.
-0.411**
-0.433**
-0.194
-0.165
Corn price vol.
0.416
0.602**
-0.443** -0.332*
Soy price vol.
-0.24
-0.236*
0.336*
0.362
Rice price vol.
Fertilizer price
-0.009
-0.029
-0.047*
-0.022
Weather expectation
0.019**
0.014
-0.009
-0.016*
Time trend
-0.001
-0.001
0.003*
0.003*
N
1130
1126
1155
1151
Test for AR(1): p-value
0.001
0.001
0.076
0.075
Test for AR(2): p-value
0.423
0.413
0.419
0.390
Note: Robust standard errors in parenthesis; * p<0.10, ** p<0.05, *** p<0.01.
Soybeans
(1)
(2)
0.922***
-0.145**
0.897***
Rice
(1)
0.682***
-0.092
-0.171*
-0.223**
0.319***
0.294**
0.065**
0.214
-0.258
0.208
0.037
0.029*
0
1100
0.007
0.235
0.164
-0.527
0.569
0.056
0.026
0.001
1096
0.006
0.241
-0.19
-0.021
0
0.002
1332
0.018
0.313
Yield response
Wheat
Variable
Own spot price
(1)
Corn
(2)
0.169***
Own futures price
(1)
Soybeans
(2)
0.05
0.166***
(1)
(2)
0.162***
0.026
Rice
(1)
0.034**
0.180**
Own price volatility
-0.483**
-0.477** -0.107
-0.1
-0.09
0.059
-0.023
Fertilizer price
-0.057***
-0.062** -0.014
0
Weather condition/expectation
0.078***
0.071*** 0.108*** 0.105*** 0.145*** 0.150*** 0.084***
Time trend
0.018***
0.019*** 0.026*** 0.026*** 0.016*** 0.016*** 0.016***
-0.042* -0.061** -0.026**
N
1147
1145
1412
1412
1340
1338
1332
Test for AR(1): p-value
0.001
0.001
0.001
0.001
0.001
0.001
0.000
Test for AR(2): p-value
0.118
0.117
0.703
0.750
0.067
0.073
0.278
Note: Robust standard errors in parentheses; * p<0.10, ** p<0.05, *** p<0.01