Agricultural Price Discovery and Forecasting in the United ... Discovery and... · –Agricultural Marketing Service (AMS), Economic Research Service (ERS), Foreign Agricultural Service

1

Agricultural Price Discovery and Forecasting in the United States

2019 India Agricultural Outlook Forum

Michael K. Adjemian

2

What I will talk about

• Factors that affect agricultural commodity prices

• How commodity prices are discovered and shared in the United States

– What role does USDA play?

• Price forecasting methods

3

Factors that affect commodity prices

• Macroeconomic shocks

– Population growth, per-capita income, exchange rates, financial shocks

4

Factors that affect commodity prices

• Seasonality of supply

– Weather fluctuations

– Production response is lagged to market signals

• Storability, trade, and transfer costs

• Derived demand

• Inelasticity of supply and demand

5

Price discovery

• Prices clear the market… if they are allowed to

• New information alters price expectations & behavior among market participants – They revise their planting/production/investment

choices

• Examples of info: poor weather forecasts, embargoes, trade war news, animal disease concerns

6

Quality of price discovery related to market structure

• Number of traders -> liquidity – Increased competition

• Transparency

– Reduce information asymmetries and potential for manipulation

• Ease of arbitrage

– Permits prices to find natural equilibria over space and time

7

Price discovery in the United States • Major grains, softs, cattle, pigs, & milk

– Futures and cash markets

• Minor commodities, poultry, and specialty crops

– Cash markets only

8

U.S. cash markets

• Found all over the nation – Major mkt prices reported by

USDA-AMS

• Immediate transfer

• Some are more competitive than others – Export terminals, processing areas

– Just two purchasers of in-shell

peanuts in Georgia

9

U.S. futures markets • Rules are meant to foster liquidity

– Homogeneous assets: quality, transaction date, location

– Only negotiable is price

– Represent just a few cash markets

• Most are highly competitive – Prices reflect all public news; they represent equilibrium

difference of opinion about expected mkt. fundamentals

• Prices discovered there used as cash mkt. benchmarks – Cash market prices often quoted in terms of basis

– Basis= Pcash – Pfutures

10

Cash - futures over space

• Avg. corn basis in August, 2016-2018

Proximity to cheap/river

transport

Dairy & poultry- feeding

operations

Cattle feedlots

Distant from demand centers

11

Market transparency • Crucial to efficiency & robustness

– Support planning and hedging / transfer of risk

– Insufficient or asymmetric information reduces liquidity and economic activity

• Three principal sources of commodity market information in the United States – Futures markets

– Private news services

– The USDA (strong reputation for objectivity) • Why? Levels the playing field, enhances transparency, and

portrays current and expected market conditions

12

How the USDA shares information • Routine, scheduled publications

– Highly anticipated by traders due to their objectivity & comprehensive nature

– Relied on for guidance about market conditions/expectations

• Department organizes, collects, and/or reviews

– Farm surveys from the National Agricultural Statistics Service (NASS)

– Trade data – Foreign attache reports – Satellite imagery – Weather forecasts – Govt. program participation

• Publishes domestic and international crop and livestock

production, supply, use, inventories, & prices

13

U.S. farm commodity data • Supply side

– Farmers surveyed for planting intentions (March), and acreage choices (June)

– Crop yield and condition is observed by field enumerators; they identify and monitor random sample plots (throughout growing season)

• Demand side

– On- and off-farm stocks are monitored via survey – Feed use depends on herd size, makeup, and location,

which are identified and reported by USDA – Food & industrial use is modeled by analysts – Export sales and inspections are reviewed

14

How the USDA prepares reports • Interagency group

– Agricultural Marketing Service (AMS), Economic Research Service (ERS), Foreign Agricultural Service (FAS), and the Farm Service Agency (FSA)

• Meet the morning ahead of publication, in sealed, “lockup” conditions

• Revise projections using… – Carefully-guarded NASS data and other information

– Economic models and statistical analysis

– Expert judgement

15

Markets react to USDA news • USDA publishes most reports at noon in the U.S.

Eastern time zone; 10:30pm in New Delhi

• Reactions can be substantial – Forward & current prices adjust to reflect new expected

supply and demand equilibrium – Options premia fall as market uncertainty about crop

conditions is reduced – Evidence that confidential preparation process works

• Markets fully absorb USDA news within a few

minutes…

16

Avg. corn futures market reaction to USDA news, 2009-2019

Scal

ed V

ola

tilit

y; 1

= “

no

rmal

”

Source: Adjemian and Irwin (2019)

17

World cotton market • Three largest producers: India (23%), China(22%),

U.S. (18%)

• Three largest users: China (33%), India (20%), Pakistan (8.7%)

• U.S. only uses 2.5% of the world cotton production; it is a net exporter

• U.S. cotton prices are affected by the price of substitutes (e.g., wool and synthetics) and regional supply and demand conditions/ trading patterns

18

Cotton price discovery in the U.S. • ICE futures No.2 contract delivers in SC/TX/TN

– Mar, May, Jul, Oct, and Dec expiries

• USDA-AMS issues cash prices & quality bulletins in 7 major production areas – Dumas, AR; Visalia, CA; Rayville, LA; Florence, SC;

Memphis, TN; Corpus Christi, TX; Lamesa, TX

– Fiber color, strength, length, etc.

• Regular USDA reports & other market developments inform prices

New crop Old crop

19

Forecasting commodity prices

• Structural supply and demand model

• Time series approach

20

Structural supply and demand model

• Inputs – Price and quantity data

– Supply and demand elasticites (flexibilities)

• Can be cumbersome – How partial/general do you want to make the

model?

• Misspecification errors can be costly

21

Simple structural model

Dcotton

S2USDA

S1USDA

Qcotton

Pcotton

ΔP

ΔUSDA Production Forecast

22

Time series approach

• Many choices – Simple vs. sophisticated

• Condition price on fundamentals – For example: year-end stocks value summarizes supply and

demand conditions – High expected stocks/expected -> lower prices – Low expected stocks/use -> higher prices – Effect is asymmetric when storage is possible

• Price forecast generated by inputting fundamentals updates – Incorporate uncertainty in forecasts?

23

Corn Prices vs Ending Stocks

24

Corn Prices vs Ending Stocks

25

Corn Prices vs Ending Stocks

*R2 increases to 73% when including a single price lag

26

Forecast Corn Prices

*Include other factors to improve forecast precision…

Model performance improves post-RFS

Except for period of very low stocks

27

Forecast Corn Prices

Source: Adjemian, Bruno, and Robe (2019)

28

Thank you!

Michael K. Adjemian

[email protected]

29

USDA Forecasting Process

Source: Liefert (2019)