Price transmission, vulnerability mapping
and early warning
Matthias Kalkuhl and Anna Winter
ZEF-IFPRI Workshop on Food Price Volatility and
Food Security, Bonn
8 July 2014
Motivation
Much analysis on price transmission, but:
• Focus only on selected countries and commodities
• Existing studies differ on data and methodology
Comprehensive analysis on market linkages and food security
implications missing (‘global picture’)
Use transmission analysis for vulnerability mapping and early
warning
Aim of this Research
Fill the gap by providing a comprehensive transmission
analysis and vulnerability mapping
• Almost universal country coverage
• Large set of international reference prices as FAO Food
Price Index might not be appropriate for all countries
• Appropriate commodity coverage which is
representative for the food basket of poor consumers
Theoretical Framework
• Standard trade model:
Transmission
elasticity (importer)
• Substitution effects from traded good j to non-traded good i:
Inelastic supply of the non-traded good i in the
short-term gives the elasticity:
Theoretical Framework
direct transmission
(traded goods)
Indirect transmission
(domestic substitution
effects to non- traded goods)
transmission through transmission through
traded substitutes
non-traded substitutes
Substitution effects on international market
Theoretical Framework
transmission
through trade
substitution
effects
transmission
through trade
substitution
effects
Domestic Market
International Market
Methodology
domestic
food price
internat
food price
exchange
rate
internat
oil price
seasonality
(monthly
dummy)
Considered Data: International
Reference Prices
63 International price indices as well as commodity prices
from major export markets and commodity exchanges
• FAO food price indices / subindices
• World Bank price indices and commodity prices
• IGC spot prices and price indices
• Futures prices USA, South Africa (Bloomberg)
• FAO Giews International Price Database
Domestic Reference Prices
• Food price indices from national CPI statistics (ILOSTAT)
• Grain price index from retail prices (FEWS.NET, WFP, FAO
GIEWS, national sources)
– Maize, wheat, rice, sorghum, millet
– Prices weighted according to domestic food supply (g/day/capita)
during 2000-2009 (FAOSTAT)
Share of grain price
index on total domestic
caloric supply
Data and Specification
Transmission to Domestic GPI:
Vulnerability Map
Number and Extent of People Affected
• Map transmission elasticities with the number of people in the country below
the extreme poverty line
• Commodity-specific Lorenz-type curves on transmission and affected population
Robustness Analyses
• Significance levels
• Deflated prices
• OLS vs. Newey-West
Transmission – Conclusions
• Domestic prices are closely linked to international prices
• ~90% of the poor in countries where international prices are
transmitted; ~400 million with transmission of >30%
• Transmission sensitive to considered price series
• Heterogeneous welfare impacts of price changes:
short-term shocks and volatility with negative welfare impacts
• Use transmission for identifying vulnerable countries (early
warning)
Paper available as:
Kalkuhl, M. (2014) "How strong do global commodity prices influence domestic
food prices in developing countries? A global price transmission and vulnerability
mapping analysis" ZEF Discussion Papers on Development Policy No.191
Early Warning System (EWS)
1. Transmission Analysis
2. Partial Price Changes
Based on historical price data
Based on current global prices
Calculation of transmission
elasticity for each country
separately
•
•
•
Estimate partial effect on
domestic price changes
(next 3 months)
3. Risk Indicator
Relative to domestic volatiliy
Alert (=red) only if predicted
domestic price changes are
abnormally high
EWS is supposed to link price dynamics from international markets directly to FNS for each country
separately.
Not just for identification of established crises but rather for short-term alerts in advance.
Target groups:
- National governments, IOs, NGOs
could react at an early stage
- Responsible financial investors
could adapt their trading activity
15
Methodology of the EWS
1. Estimate Transmission from International Prices
Methodology of the EWS
2. Estimating Partial Price Changes
Methodology of the EWS
Illustration of the EWS
Illustration:
EWS for FAO Food Price Index
0
5
10
15
Number of countries
with alerts:
2007m1
2008m1
upper CI/lower CI
2009m1
t
2010m1
# countries with alerts
2011m1
15
10
5
0
Illustration:
EWS for
FAO Food Price Index
2008m1
15
2007m1
Number of countries
with alerts:
2009m1
t
0
5
10
upper CI/lower CI
2000m1
2002m1
2004m1
2006m1
t
upper CI/lower CI
2008m1
2010m1
2012m1
# countries with alerts
2010m1
# countries with alerts
2011m1
Illustration: Countries with Alerts
No. of
countries
with alerts
No. of
countries
with alerts
No. of
countries
with alerts
No. of
countries
with alerts
Illustration: No. of People Affected
people affected
(in m.)
people affected
(in m.)
Illustration: EWS as Excel Tool
Excel tool as a user
friendly instrument for
both policy makers and
financial investors:
Properties of the tool:
-
Alert parameters
(thresholds, confidence
intervals) can be adjusted
-
Shows predicted partial
price changes on domestic
markets for the next three
months at different
confidence intervals by
country and prices (indices)
-
Estimates the expected
number of affected people
and countries at different
confidence intervals
Discussion
Thank you for your attention!
www.zef.de/volatility.html
Appendix
Number and Extent of People Affected
• Using aggregated price indices neglects the heterogeneity of international prices
that are relevant for particular countries due to heterogeneous trade patterns
• FAO Food Price Index can nevertheless give a broad indication
Number and Extent of People Affected
• Different Poverty Lines
Transmission vs. Pass-through
Wheat Prices
Corn Prices
Thai Rice Prices
Rice Prices
International Reference Prices
w
1
Variable
FAO food price index
Source
FAO
FAO cereals price index
FAO oil/fat price index
Description
Consists of 55 commodity quotations considered as representing the international prices of food commodities; weighted by
export share
Consists of wheat, maize and rice prices; weighted by export share
Consists of 12 different oils (including animal and fish oils); weighted by export share
2
3
4
FAO sugars price index
Index form of the International Sugar Agreement prices with 2002-2004 as base
FAO
5
FAO meat price index
Consists of poultry, bovine meat, pig meat and ovine meat products; weighted by export share
FAO
6
FAO diary price index
Consists of butter, skimmed milk powder, whole milk powder, cheese and casein prices; weighted by export share
FAO
7
8
WB grains price index
WB fats and oils price index
Includes barley, maize, rice and wheat
Includes coconut oil, groundnut oil, palm oil, soybeans, soybean oil and soybean meal.
World Bank
World Bank
9
Wheat (HRW) US
No. 1, hard red winter, ordinary protein, export price delivered at the US Gulf port for prompt or 30 days shipment
World Bank
10
Wheat (SRW) US
No. 2, soft red winter, export price delivered at the US Gulf port for prompt or 30 days shipment
World Bank
11
Wheat CAN
Wheat (Canada), no. 1, Western Red Spring (CWRS), in store, St. Lawrence, export price
World Bank
12
Wheat AUS
Australian soft white, Australia, f.o.b.
Australia Eastern States Standard White Wheat FOB Spot (for 10/2007-09/2008 where USDA/IGC series has missing entries)
USDA/IGC
Bloomberg
13
Barley
Barley (Canada), feed, Western No. 1, Winnipeg Commodity Exchange, spot, wholesale farmers' price
World Bank
14
Sorghum US
Sorghum (US), no. 2 milo yellow, f.o.b. Gulf ports
World Bank
15
Corn US
Maize (US), no. 2, yellow, f.o.b. US Gulf ports
World Bank
16
Soybeans
Soybeans (US), c.i.f. Rotterdam
World Bank
17
Soybean oil
Soybean oil (Any origin), crude, f.o.b. ex-mill Netherlands
World Bank
18
Soybean meal
Soybean meal (any origin), Argentine 45/46% extraction, c.i.f. Rotterdam beginning 1990; previously US 44%
World Bank
19
Rice Thai A1
World Bank
20
Rice Thai 5%
21
Rice Thai 25%
Rice (Thailand), 100% broken, A.1 Super from 2006 onwards, government standard, f.o.b. Bangkok; prior to 2006, A1 Special,
a slightly lower grade than A1 Super
Rice (Thailand), 5% broken, white rice (WR), milled, indicative price based on weekly surveys of export transactions,
government standard, f.o.b. Bangkok
Rice (Thailand), 25% broken, WR, milled indicative survey price, government standard, f.o.b. Bangkok
22
Rice Vietnam
Vietnamese rice, 5% broken
World Bank
23
Palm oil
Palm oil (Malaysia), 5% bulk, c.i.f. N. W. Europe
World Bank
24
Groundnut oil
Groundnut oil (any origin), c.i.f. Rotterdam
World Bank
25
Coconut oil
Coconut oil (Philippines/Indonesia), bulk, c.i.f. Rotterdam
World Bank
26
Fishmeal
Fishmeal (any origin), 64-65%, c&f Bremen, estimates based on wholesale price, beginning 2004; previously c&f Hamburg
World Bank
27
Beef
Meat, beef (Australia/New Zealand), chucks and cow forequarters, frozen boneless, 85% chemical lean, c.i.f. U.S. port (East
Coast), ex-dock, beginning November 2002; previously cow forequarters
World Bank
FAO
FAO
World Bank
World Bank
35
International Reference Prices
w
28
Variable
Chicken
Source
World Bank
Wheat / CBT
Description
Meat, chicken (US), broiler/fryer, whole birds, 2-1/2 to 3 pounds, USDA grade "A", ice-packed, Georgia Dock preliminary weighted
average, wholesale
Meat, sheep (New Zealand), frozen whole carcasses Prime Medium (PM) wholesale, Smithfield, London beginning January 2006;
previously Prime Light (PL)
#2 Soft Red Winter at contract price, #1 Soft Red Winter at a 3 cent premium, Chicago Board of Trade
29
Sheep
30
31
Corn / CBT
#2 Yellow at contract Price, #1 Yellow at a 1.5 cent/bushel premium #3 Yellow at a 1.5 cent/bushel discount, Chicago Board of Trade
Bloomberg
32
Soybeans / CBT
#2 Yellow at contract price, #1 Yellow at a 6 cent/bushel premium, #3 Yellow at a 6 cent/bushel discount, Chicago Board of Trade
Bloomberg
33
Soybean oil / CBT
Crude soybean oil meeting exchange-approved grades and standards, Chicago Board of Trade
Bloomberg
34
35
Soybean meal / CBT
Rough Rice / CBT
Bloomberg
Bloomberg
36
Feeder Cattle / CME
48% Protein Soybean Meal, Chicago Board of Trade
U.S. No. 2 or better long grain rough rice with a total milling yield of not less than 65% including head rice of not less than 48%,
Chicago Board of Trade
650-849 pound steers, medium-large #1 and medium-large #1-2, Chicago Mercantile Exchange
37
Live Cattle / CME
55% Choice, 45% Select, Yield Grade 3 live steers, Chicago Mercantile Exchange
Bloomberg
38
Lean Hogs / CME
Hog (barrow and gilt) carcasses, Chicago Mercantile Exchange
Bloomberg
39
Wheat / KCBT
Hard Red Winter Wheat, No. 2 at contract price; No. 1 at a 1 1/2-cent premium; Kansas City Board of Trade
Bloomberg
40
Wheat / MGEX
Bloomberg
41
White Maize / SAFEX
Hard Red Spring Wheat, No. 2 or better Northern Spring Wheat with a protein content of 13.5% or higher; Minneapolis Grain
Exchange
South African Futures Exchange; starting in 08/1996
42
Yellow Maize / SAFEX
South African Futures Exchange; starting in 08/1996
Bloomberg
43
Wheat / SAFEX
South African Futures Exchange; starting in 11/1997
Bloomberg
44
Soybean / SAFEX
South African Futures Exchange; starting in 04/2002
Bloomberg
45
South African Futures Exchange; starting in 02/1999
Bloomberg
46
Sunflower Seeds /
SAFEX
Palm oil / MDEX
Malaysia Derivatives Exchange; starting in 03/1995
Bloomberg
47
GSCI Agriculture
Own calculation
48
Unweighted average over soybean, soybean oil, soybean meal, lean hog, feeder cattle and live cattle futures
Own calculation
50
US Cereals Futures
Index
US Oils & Meats
Future Index
US Futures Index
Price index over active futures with the 2012 S&P GSCI weights on wheat (CBT), wheat (KCBT), corn, soybeans, lean hogs, live cattle
and feeder cattle (all CBT)
Unweighted average over wheat, corn and rice futures
Average of US Cereals and US Oils & Meats Future Index
Own calculation
51
Non-US Futures Index
Average over non-US futures (SAFEX maize and wheat, MDEX palm oil)
Own calculation
52
All Futures Index
Average of US Futures and non-US Futures Index
Own calculation
49
World Bank
Bloomberg
Bloomberg
Bloomberg
Own calculation
36
Theoretical Framework
Theoretical Framework
Domestic
Storage
yes/no
Global
Storage
yes/no
none, but
expected
yes
yes
none, but
expected
yes
no
yes
none, but
no
expected
none and
yes
not expected
none and
no
not expected
yes/no
yes/no
yes/no
0
Volatility and Underweight
(1)
(2)
(3)
(4)
(5)
(6)
-0.309***
(-2.42)
-0.336***
(-2.55)
-0.181
(-1.51)
-0.325**
(-2.59)
-0.315**
(-2.49)
-0.273**
(-2.01)
L. Log GDP per
capita (PPP)
-0.240
(-1.36)
-0.245
(-1.38)
-0.322*
(-1.87)
-0.240
(-1.42)
-0.242
(-1.43)
-0.323*
(-1.73)
Log Sanitation
-0.344*
(-1.95)
-0.344*
(-1.88)
-0.286*
(-1.70)
-0.333*
(-1.90)
-0.347**
(-2.00)
-0.333*
(-1.86)
Food price volatility
0.011**
(1.98)
0.004
(0.74)
0.027***
(2.60)
0.018*
(1.77)
0.029**
(2.45)
0.006
(0.63)
L. Food price volatility
0.017***
(2.61)
0.013**
(1.99)
0.034***
(2.74)
0.018*
(1.76)
0.016
(1.42)
0.004
(0.35)
0.033
1264
92
0.085
1263
92
0.014
1326
96
0.146
1325
96
0.033
1326
96
0.805
1322
96
Log GDP per
capita (PPP)
P-values joint significance of
volatility and lag volatility
Observations
Number of countries
Note. Dependent variable: log of prevalence of underweight for children under five years. z-values based on bootstrapping with
1000 replications in parentheses. *** p<0.01, ** p<0.05, * p<0.1. Constants are omitted.
Volatility and Stunting
(1)
(2)
(3)
(4)
(5)
(6)
-0.302***
(-2.67)
-0.304***
(-2.74)
-0.272***
(-2.59)
-0.321***
(-2.81)
-0.328***
(-2.85)
-0.270**
(-2.34)
Lagged log GDP per
capita (PPP)
-0.121
(-0.89)
-0.151
(-1.12)
-0.16
(-1.27)
-0.135
(-1.07)
-0.126
(-0.94)
-0.218
(-1.51)
Log Sanitation
-0.226
(-1.63)
-0.221
(-1.64)
-0.211
(-1.62)
-0.226*
(-1.76)
-0.236*
(-1.80)
-0.219
(-1.59)
Log volatility
0.009**
(2.03)
0.009
(1.57)
0.011*
(1.96)
0.013
(1.63)
0.009
(1.27)
0.011
(1.32)
Lagged log volatility
0.017***
(3.6)
0.017***
(3.44)
0.013**
(2.19)
0.011
(1.56)
0.007
(1.06)
0.002
(0.34)
P-values joint significance of
volatility and lag volatility
0.001
0.001
0.079
0.239
0.426
0.294
Observations
Number of countries
1256
92
1255
92
1304
96
1303
96
1304
96
1300
96
Log GDP per
capita (PPP)
Note. Dependent variable: log of prevalence of stunting for children under five years. z-values based on bootstrapping with 1000
replications in parentheses. *** p<0.01, ** p<0.05, * p<0.1. Constants are omitted.
Volatility Measures
Illustration: Countries with Alerts
No. of
countries
with alerts
No. of
countries
with alerts
No. of
countries
with alerts
No. of
countries
with alerts
Overview information tools