Title: Real rate of production : the income and insurance effects of agricultural policy
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00089823/00001
 Material Information
Title: Real rate of production : the income and insurance effects of agricultural policy
Abbreviated Title: Journal reprint series - Internation Agricultural Trade and Policy Center ; 05-04
Physical Description: Book
Language: English
Creator: Thompson, Stanley R.
Schmitz, P. Michael
Iwai, Nobuyuki
Goodwin, Barry K.
Publisher: Institute of Food and Agricultural Sciences, University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: August, 2005
 Record Information
Bibliographic ID: UF00089823
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.

Downloads

This item has the following downloads:

PDF ( PDF )


Full Text

JRTC 05-04


I '-ional Agricultural Trade and Policy Center




THE REAL RATE OF PROTECTION: THE INCOME AND
INSURANCE EFFECTS OF AGRICULTURAL POLICY
By
Stanley R. Thompson, P. Michael Schmitz, Nobuyuki Iwai, & Barry K.
Goodwin
JRTC 05-04 August 2005


JOURNAL REPRINT SERIES


'V


i~fr


UNIVERSITY OF
FLORIDA


Institute of Food and Agricultural Sciences









INTERNATIONAL AGRICULTURAL TRADE AND POLICY CENTER


THE INTERNATIONAL AGRICULTURAL TRADE AND POLICY CENTER
(IATPC)

The International Agricultural Trade and Policy Center (IATPC) was established in 1990
in the Institute of Food and Agriculture Sciences (IFAS) at the University of Florida
(UF). The mission of the Center is to conduct a multi-disciplinary research, education and
outreach program with a major focus on issues that influence competitiveness of specialty
crop agriculture in support of consumers, industry, resource owners and policy makers.
The Center facilitates collaborative research, education and outreach programs across
colleges of the university, with other universities and with state, national and
international organizations. The Center's objectives are to:

* Serve as the University-wide focal point for research on international trade,
domestic and foreign legal and policy issues influencing specialty crop agriculture.
* Support initiatives that enable a better understanding of state, U.S. and international
policy issues impacting the competitiveness of specialty crops locally, nationally,
and internationally.
* Serve as a nation-wide resource for research on public policy issues concerning
specialty crops.
* Disseminate research results to, and interact with, policymakers; research, business,
industry, and resource groups; and state, federal, and international agencies to
facilitate the policy debate on specialty crop issues.









Applied Economics, 2004, 36, 1-8


The real rate of protection: the income and

insurance effects of agricultural policy


STANLEY R. THOMPSON*t, P. MICHAEL SCHMITZ$,
NOBUYUKI IWAIt and BARRY K. GOODWINt
tDepartment of Agricultural, Environmental, and Development Economics,
Bm 335 Agricultural Administration Bildi;:, The Ohio state university,
2120 Furse Road, Columbia, Ohio 43210, USA and lInstitute of Agricultural Policy
and Market Research at the University of Giessen, Giessen, Germany




Agricultural price policies in developed countries aim at protecting farmers against
both low and volatile world market prices. However, traditional indicators of pro-
tection only refer to the income (level) effect of policy. Following other research, it is
argued that public policy can also yield an insurance (stabilizing) effect. In this paper
a way to measure these dual effects is proposed. The method is illustrated with wheat
market data for the USA and the European Union. Strong evidence is found that the
insurance effect is an important component of protection, albeit a small one relative
to the income effect. Policy support provided higher income and lower insurance
effects in the EU than in the USA. For both markets, policy reforms in the 1990s
led to significantly reduced income effects and smaller insurance effects. Without
accounting for the influence of policy on income variability, traditional measures
of protection will understate the real rate of protection.


I. INTRODUCTION

Countries protect their agricultural sectors for many differ-
ent reasons and in many different ways. Protectionist poli-
cies are often entrenched in long histories of political and
economic compromise. As a result they are often multifa-
ceted and complex. However, two basic types of policy sup-
port to agriculture are generally used: market price supports
and budgetary payments. Market price supports often take
the form of price interventions (including export subsidies
and import tariffs) and payments based on market prices
and output (deficiency payments). On the other hand, bud-
getary payments take the form of direct payments to pro-
ducers based on area planted, historical entitlements and the
like. Because budgetary payments are generally not directly
tied to production decisions, they are often considered to
be decoupled. While all these measures are generally
thought of as supporting producer incomes, in a world of
uncertainty additional risk (insurance) benefits may result.

*Corresponding author. E-mail: thompson.sl@osu.edu


Hennessy (1998) examined the wealth (income) and
insurance (risk) effects of policy on optimal production
decisions. He argued that studies of trade policy reform
in stochastic environments should consider both insurance
and income effects. In this paper, these two components
are examined, however, within a different context: the mea-
surement of the degree of policy-induced protection. A
method to evaluate both the income and insurance compo-
nents of protection is proposed. The impact of major policy
reform on the income and insurance effects is assessed.
Among other things, insight is offered into the influence
of coupled and decoupled policies on the two components
of protection.
Traditional indicators of protection rates only refer to
the income (level) effect of policy. For instance, the nom-
inal protection rate simply indicates the percentage by
which the domestic price exceeds the border price, the effec-
tive protection rate incorporates a value-added dimension
while the producer support estimate (PSE) is yet a more


Applied Economics ISSN 0003-6846 print/ISSN 1466-4283 online ( 2004 Taylor & Francis Ltd
http://ww 1 .i. 11 i i . ~i. .
DOI: 10.1080/0003684032000164102


[1-8] [Page No. 1] {Tandf}Raef/RAEF-100942.3d


R out1iedge


[25.11.2003-8:53pm] [FIRST PROOFS]


(RAEF) Paper: RAEF-100942









S. R. Thompson et al.


comprehensive measure in that it attempts to account for
all domestic policy transfers to producers. It is argued that
these measures by themselves are incomplete since they
ignore the income stabilizing or insurance effect.1
In fact, agricultural price policies in developed countries
aim at protecting farmers against both low and volatile
world market prices. It is argued that this kind of 'double
protection' can only be computed on the basis of an
expected utility approach, measuring the percentage
increase in the expected utility of either income or of the
certainty equivalent of income, respectively (Schmitz, 1997).
Following Newbery and Stiglitz (1981), a mean-coefficient
of variation approximation with log-normally distributed
incomes is used to derive an expression that is termed the
real rate of protection (RRP). The proposed measure is
illustrated with wheat market data from the USA and
European Union. While it is not the intention to provide
explicit measures of the effects of actual policy instruments,
it is illustrated how the incorporation of insurance effects
might affect standard measures. Thus, the analysis is based
upon a number of assumptions regarding the nature of pre-
ferences, the relevant market prices and risk measures, and
other parameters inherent in the analysis.


II. METHODOLOGY

An expected utility approach that incorporates the first and
second moments of the probability distribution of income
is applied. Assuming a log-normal distribution of real
income and constant relative risk aversion utility, the
mean-coefficient of variation formula of the expected
utility of income is2

E[U(y)] = E[y][1 +J 0.5 (1)
where, U= utility, y= income, cvy =coefficient of variation
of income and, R = coefficient of relative risk aversion.
Defining the real rate of protection as the percentage
change in the expected utility of income with protection (Ki)
vis-a-vis without protection (%o), the real rate of protection
(RRP) can be derived:3

RRPi = (1 + ERPi) 1 )0.5 1 (2)

where ERPi= effective rate of protection (per cent change
in income) and ci, =c. .nccicn of variation of income:
i=0 without protection, i 1 with only price and output
protection, and i= 2 with all protection.


In the international trade literature, the effective protec-
tion coefficient (EPC) is conventionally defined as the ratio
of actual value added by domestic resources to the value
added in a market free of distortions. The effective pro-
tection rate follows simply as 100(EPC-1). A similar
value added measure of income (ERP) is used, defined as
(Yi- Yo)/Yo, i= 1,2, where income is defined as revenue
less variable cost and the subscript i refers to different
degrees of policy protection.
In the assessment of how policy impacts the coefficient
of variation of income, two scenarios are considered, each
differing by the assumptions made about the probability
distribution of price and quantity. First, it is assumed
that the coefficient of variation of income is determined
by the single random variable price and its variance is
constant over the sample period. Given adaptive expecta-
tions and market rigidities this scenario posits quantity
produced to be predetermined.
In the second scenario, the assumption of constant price
variance is relaxed and time varying variances computed
using implied volatilities. Annual implied volatilities, cal-
culated from wheat futures prices on the Chicago Board of
Trade (CBOT), are used to assess the variability of wheat
prices. These volatilities allow relaxation of the assumption
that price variance is constant over time. Indeed, an exam-
ination of the implied volatilities demonstrates the fact
that market price volatilities vary considerably from year
to year, depending on market conditions. The implied vola-
tilities are the implied volatilities taken from the Bridge
database. In particular, these prices were calculated using
September prices for the following July Chicago Board of
Trade futures contracts. Using the log-normality assump-
tion, a simulated series of log-normal prices were derived
for each market. It is assumed that annual border prices
represent the mean price and that the volatility faced at the
border in each market is the same as that implied by the
CBOT options market. Thus, a series of log-normally dis-
tributed simulated prices are generated (1000 replications
are used) that display the variances implied by the CBOT
price volatility.
It is further assumed that domestic policies truncate this
distribution, such that expected prices and price variability
are altered according to the truncated distribution. In par-
ticular, the truncation point is given by the domestic sup-
port price, which serves to establish a minimum expected
price. This truncated distribution is used to calculate the
coefficient of variation of incomes under the domestic poli-
cies. This computed coefficient of variation reflects the fact


1Any measure of policy effects is likely to suffer shortcomings and the validity of any given measure depends upon exactly what aspect of
support the instrument intends to represent. Measures of market distortions may imply a quite different picture of support when
compared to measures of producer income effects, such as the one considered here.
2If it is alternatively assumed that real income is normally distributed and constant absolute risk aversion utility, the simple mean-
variance formula of expected utility of income is: E [U (y)] = E(y) 1/2 A var(y) where y = income and, A = coefficient of absolute risk
aversion.
3Equation 2 is derived in the Appendix.


[1-8] [Page No. 2] {Tandf}Raef/RAEF-100942.3d


[25.11.2003-8:53pm] [FIRST PROOFS]


(RAEF) Paper: RAEF-100942









Income and insurance effects of agricultural policy

that stochastic prices are stabilized by policies undertaken
to protect domestic markets from international effects.
The second scenario also accounts for stochastic quan-
tities. Specifically, crop yields are assumed to be log-
normally distributed with simple price-yield correlations
simulated over reasonable negative values.4 The method
of Johnson and Tenebein (1981) is then used to draw cor-
related random yields and prices from which is generated
a series of log-normally distributed yields. The assumptions
of this scenario allow for the explicit examination of policy
regime change. This is accomplished through subperiod
analysis.


III. EMPIRICAL ILLUSTRATIONS -THE
CASE OF US AND EUROPEAN WHEAT

The data
Annual data for 1986-2000 were used to implement the
theoretical model. Wheat market data for the USA and
EU provide the empirical illustrations. Producer prices,
border prices, quantities produced and budgetary pay-
ments were obtained from the Organization for
Economic Cooperation and Development (OECD various
issues). Annual data on variable production costs per hec-
tare for wheat over this same period were obtained from
the US Department of Agriculture and Eurostat (2001).
All data were available for the EU-15, treated as a single
entity, however input costs are available only for individual
countries. Since wheat production costs can vary widely
across the EU, it was chosen to proxy the EU with vari-
able cost data for France. By volume, France is the largest
wheat producer in the EU. Incomes are defined in per hec-
tare terms. All data are available from the authors upon
request.


Policy environment
Over the last two decades in the European Union, the 1992
MacSharry reform and the Agenda 2000 have been the
major policy reforms to the EU's Common Agricultural
Policy. The Agenda 2000 continues and deepens the
MacSharry reforms. The two basic policy instruments
used for cereals are: market price support (MPS) which
includes the use of administered prices, export subsidies,
and import tariffs and; area payments which are direct
payments based on current acreage of the 'grand cultures'
(cereal, oilseeds, protein plants) conditional subject to
land set-aside requirements. MPSs are considered to be
'coupled' in the sense that they are directly production


and trade distorting, while direct payments are widely
accepted as less distorting and more targeted to the income
objects. The transition from MPS to direct payments under
MacSharry was announced as an effort to decouple pay-
ments to producers, although this is not quite true since
the payments are still coupled to the factors of production.
Since 1992 direct (compensatory) payments have been
made to partially compensate for the market distorting
MPSs. Indeed, a reduction in MPSs has continued follow-
ing Agenda 2000 but area payments have increased.
Important differences in the price and producer welfare
effects of policy have been found between the pre- and
post-MacSharry regimes (Thompson et al., 2000; Thompson
et al., 2002a; Thompson et al., 2002b). The ex post analysis
does not include the period of Agenda 2000.
The USA has also moved to decouple supports with
direct income transfers instead of interventions that place
prices above market clearing levels. The 1996 FAIR Act
phased out the deficiency payments and acreage set-asides
of the 1990 farm bill in favour of fixed decoupled pay-
ments, termed production flexibility contracts payments
(PFCs). PFC payments are historically based and widely
considered to be decoupled. Since 1996, however, addi-
tional producer support programmes have been added.
For example, in 1999 PFC payments were supplemented
with a sizeable 'market loss assistance' programme and in
2001, with a costly crop insurance premium subsidy. While
1996 reforms were a major step toward a more market
oriented agriculture, ad hoc additions have hampered lib-
eralization progress. Provisions of the 2002 Farm Bill
became effective after the sample period used for the
empirical illustrations. Like the EU, two major policy
regime periods for the USA are identified as pre- and
post-FAIR.


IV. MODEL IMPLEMENTATION


To empirically illustrate the analytical framework, farmer
income estimates are derived both with and without policy,
the latter measure being the situation that would have
occurred if there were no policy distortions. The former
estimate considers two policy scenarios: producer incentive
supports (actual market price and output payments) and
total support (support including the addition of direct
payments). Thus, three measures of producer incomes are
computed:
without support:

Y = [Pb (Pp)]- VC (3)


4The exact value used for the Spearman correlation coefficient is -0.4. This value was used for corn prices by Hennessy et al. (1997). They
based their calculation on test plot data for Iowa corn markets. Of course, this value may vary from year to year and certainly may
be different across alternative crops, including wheat. This coefficient was found to be a reasonable estimate as our simulations were
relatively insensitive to alternative values.


[1-8] [Page No. 3] {Tandf}Raef/RAEF-100942.3d


[25.11.2003-8:53pm] [FIRST PROOFS]


(RAEF) Paper: RAEF-100942









S. R. Thompson et al.


with producer incentive support:

Yi = [P, Q(P) + Po] VC (4)
with total support:

Y2 = [Pp Q(Pp) + Bp] VC
where, Pb is the border price, Pp is the producer price, Q is
the quantity produced, P0 is output (deficiency) payments,
Bp is total budgetary payments (including Po) and, VC is
variable cost of production. All quantities and payments
are on a per hectare basis. Using these definitions, Yo is the
income that would have occurred in a well-functioning
market free of policy distortions; Y1 is the income resulting
from incentive (actual market price and output) supports
and, Y2 is the income with incentive supports plus all
other budgetary payments to producers. Producer incentive
supports are considered to be coupled to production (thus
influencing production and trade) while the additional
budgetary payments (area payments and payments based
on historical entitlements) are considered effectively
decoupled.
Descriptive statistics for price and incomes are shown in
Table 1. For the USA, producer incentive supports (market
price and output payments) increased the mean price by
15% and lowered its variablity by nearly 30%, relative to
what would have occurred without support. Incentive
supports shifted the price distribution to the right and


narrowed it significantly in dispersion. As measured by
the coefficient of variation (CV), the reduction in price
volatility was more than one-third. For the EU, producers'
incentive supports increased the mean price by more than
40% while the CV fell to less than half of the without
support scenario.
More pronounced relationships hold for producer
income. For the USA, producer incentive supports reduced
the CV of income by more than 50% with little further
reduction achieved when all support is included. For the
EU, producer incentives reduced the CV of income by 38%
percent. However, the addition of other budgetary support
reduced the CV by another 52% percent. When all support
is included, the CV of income fell by 55% and 70% for the
USA and EU, respectively. Clearly, policy interventions in
both markets increased the mean and lowered the vari-
ablility of producer incomes. The reduction in price and
income variability (risk) due to policy is an insurance effect.
Policy had a greater insurance effect on producer income
than it did on prices.


Empirical income and insurance effects
The income effect of policy is described by the effective rate
of protection (ERP). ERP1 and ERP2 are respectively, the
percentage change in mean incomes Y1 and Y2, relative to
Yo. For the USA and EU these are shown in Tables 2 and 3.


Table 1. Wheat price and producer income variation, USA and EU, 1986-2000

USA ($) EU ()
Mean Std.dev. CVa Mean Std.dev. CV
Price
Without support (PB) 106.4 28.3 0.27 108.2 21.0 0.19
With incentive support (P, + Po) 122.9 20.6 0.17 154.2 13.9 0.09
Income
Without support (Yo) 85.0 40.0 0.47 371.7 139.0 0.37
With incentive support (Y1) 116.7 25.6 0.22 657.9 151.1 0.23
With all support (Y2) 212.0 45.1 0.21 974.8 104.0 0.11
Note: acv is the coefficient of variation.


Table 2. Estimated rates of protection, n. (ERP) and real (RRP), and insurance . (IE) with constant price
variance, USA, 1986-2000

Effective rate of protection (ERP)
ERP1 0.37
ERP2 1.50
Real rate of protection (RRP)
Coefficient of relative risk aversion (R)
1.0 2.0 3.0

RRP1 0.48 0.60 0.73
IE1 (%) 8 17 26
RRP2 1.70 1.92 2.15
IE2 (%) 8 17 26


[1-8] [Page No. 4] {Tandf}Raef/RAEF-100942.3d


[25.11.2003-8:53pm] [FIRST PROOFS]


(RAEF) Paper: RAEF-100942









Income and insurance effects of agricultural policy


Table 3. Estimated rates ofprotection, n. .. (ERP) and real (RRP), and insurance n. .* (IE) with constant
price variance, European Union (15), 1986-2000

Effective rate of protection (ERP)
ERP1 0.77
ERP2 1.62
Real rate of protection (RRP)
Coefficient of relative risk aversion (R)
1.0 2.0 3.0

RRP1 0.84 0.92 0.99
IE1 (%) 4 8 12
RRP2 1.78 1.96 2.14
IE2 (%) 6 13 20


Recall, the ERPs show the percentage change in producer
income per hectare due to policy intervention.
For the USA, the value 0.37 (ERP1) implies that, on
average, incomes were 37% greater with producer incentive
supports than without protection. Similarly, the value 1.50
(ERP2) implies that incomes were 150% greater with all
supports than with no policy interventions. Clearly, over
this 15-year period, US wheat farmer incomes benefitted
greatly from farm policy interventions.
For the EU, incomes were 77% and 162% greater due to
producer incentive supports and to all support, respectively,
than would have obtained without the CAP or other policy
support structures. While the ERPs for both countries are of
the same magnitude for the all support situation, the more
market distorting producer incentive component (ERP1)
plays a much larger role for the EU than it does for the
USA. These results would suggest that, over the 1986
2000 period, EU policy instruments were relatively more
market distorting than those used by the USA.
The insurance effect enables the real rate of protection
(RRP) to be defined. It is a measure of the benefit accruing
to risk averse producers due to stabilized incomes. The
insurance effect is accounted for by the term in brackets
in Equation 2, which is composed of two determinants.
The first determinant is the relative variablility of income
as measured by the CVs. Referring back to Table 1, it was
seen that the CV(Yo)> CV(Y1)> CV(Y). Thus, for a
given degree of farmer risk aversion, R, the term contained
in brackets in Equation 2 is greater than 1.0. This means
that the variation of income without protection is greater
than with protection. If this is the situation, both incentive
supports and all budgetary payments lead to dampened
income volatility. In the special case, where policy had no
impact on income stability, the value of the insurance com-
ponent would be 1.0. A similar result would occur for the
risk neutral producer (R = 0).
Secondly, the size of the insurance effect is related to the
extent to which the producer is averse to risk (Saha et al.,
1994). The greater R, the greater the term in brackets;
hence, the greater RRP. The more risk averse the producer,


the greater the benefit received from decreased income
volatility. As farmers place a greater value on income sta-
bility, the greater the 'real' benefit they receive from policy
interventions. In other words, producer utility increases as
income volatility decreases. Although the traditional way
of evaluating price and income support policies takes the
expected value of income into consideration, it does not
consider the income volatility dimension, or what is termed
the insurance effect. The traditional approach is only valid
if producers are risk neutral or if policy has no effect on the
volatility of income.
The empirical RRP computations are shown in the bot-
tom portions of Tables 2 and 3. As expected, the RRP is
greater than the ERP and positively related to the degree of
risk aversion. The insurance effect is empirically important.
For both the USA and the EU, since income volatility was
greater without protection than under either of the two
policy interventions, the real level of wheat industry pro-
tection would be understated by traditional measures of
protection.
For the USA, the RRP1 (R = 2.0) is 0.60 while the ERPI
is 0.37; the difference between 1.60 and 1.37 is the insurance
effect. This is the percentage change in the utility of income
due to the stabilizing effect of producer incentive support
policies. In this case the stabilizing effect yields real rates of
protection that are 17% greater than would occur if policy
had not stabilized incomes. In other words, it is the value
of the policy's risk reduction effect to risk averse producers.
If the policy did not yield this insurance effect, producers
would clearly not receive this added benefit. While impor-
tant, the insurance effect remains small relative to the
income effect. Only for very risk averse producers do the
stabilizing benefits of producer incentive policies approach
the magnitude of the income effect. Empirically, the insur-
ance effect is invariant to the addition of all remaining
budgetary payments.
For the EU, the insurance effect of producer incentive
policies is about half as large as that for the USA. When
all budgetary support is included the EU insurance effect
remains relatively smaller. What stands out, is the relatively


[1-8] [Page No. 5] {Tandf}Raef/RAEF-100942.3d


[25.11.2003-8:53pm] [FIRST PROOFS]


(RAEF) Paper: RAEF-100942









S. R. Thompson et al.


larger insurance effect associated with US producer
incentive policies than those of the EU.


Policy regime analysis with stochastic yields

The impact of both price and yield variability on producer
revenue, and, hence income, depends on Cov(P,Q). A large
negative correlation would tend to stabilize incomes. The
effect of stochastic prices and yields on the ERPs is as
follows. Holding E[Yi] constant, if E[Y0] decreases, ERP,
will increase. On the other hand, holding E[Y0] constant, if
E[Yi] decreases (increases), ERPi will decrease (increase).
The magnitude of these changes is an empirical question
which we explore below.
Relaxation of the earlier constant price variance assump-
tion enables a more targeted investigation of policy regime
change. Regimes 1 and 2 are defined to represent major
policy regime periods. For the USA, this is pre- and post-
FAIR while for the EU this is pre- and post-MacSharrry.
The descriptive statistics for price and incomes are disag-
gregated by policy regime period and reported in Table 4.
For both markets, the incentive support policy changes
resulting from period 2 regime reforms led to lower price
levels and increased price variability.
The FAIR incentive support reforms resulted in increased
incomes and reduced variability. With the addition of all
other budgetary support (largely PFC payments) income
levels nearly doubled with little change in variability. The
MacSharry reforms showed reduced income levels with


incentive support reforms. However, with the addition of
all other budgetary payments (largely area or compensa-
tory payments) incomes nearly doubled during the post-
MacSharry period while income variability changed little.
The computed income and insurance effects for each policy
regime are discussed next.


Income effect. For the USA, large differences in the ERPs
(income effects) are observed between the two periods
(table 5). In regime 1, ERPI = 0.96 suggests that on average
incomes were 96% greater due to producer incentive sup-
ports. Following FAIR reforms (regime 2), ERP1 fell to 0.36.
ERP2 = 2.71 suggests that incomes were 271% greater with
all policy support than they would have been without sup-
port. Although ERP2 fell to 1.67 in the post-FAIR period,
a still somewhat high degree of overall protection remained.
For the EU, large differences in the ERPs are also
observed between the two regimes (Table 6). In regime 1,
ERP = 1.88 suggests that on average incomes were 188%
greater with producer incentive supports than without.
Following the MacSharry reforms ERP1 fell to 0.45.
ERP2=2.17 suggests that incomes were 217% greater
with all policy support than they would have been without
support. Although, RRP2 fell to 1.78 in the post-MacSharry
period, a relative high degree of overall protection remained.
For both markets, producer income effects were substan-
tially higher during regime than during the post-reform
regime period. While the income effects fell dramatically


Table 4. Wheat price and income variation by policy regime periods, USA and EU, 1986-2000

USA ($) EU ()
Regime 1 Regime 2 Regime 1 Regime 2
1986-1995 1996-2000 1986-1992 1993-2000
Price:
Without support (PB)
Mean 102.8 113.8 101.3 114.1
Std. dev. 29.3 27.7 20.0 22.2
CV 0.29 0.24 0.20 0.19
With incentive support (P, + Po)
Mean 123.6 121.4 180.3 131.4
Std. dev. 21.1 21.9 10.3 16.3
CV 0.17 0.18 0.06 0.12
Income:
Without support (Yo)
Mean 73.4 108.2 318.6 418.2
Std. dev. 43.4 18.6 135.5 132.7
CV 0.59 0.17 0.43 0.32
With incentive support (Yi)
Mean 112.1 125.9 797.6 535.6
Std. dev. 30.3 8.7 28.3 91.3
CV 0.27 0.07 0.04 0.17
With all support (Y2)
Mean 190.2 255.7 881.3 1056.5
Std. dev. 26.1 44.8 49.0 56.5
CV 0.14 0.18 0.06 0.05


[1-8] [Page No. 6] {Tandf}Raef/RAEF-100942.3d


[25.11.2003-8:53pm] [FIRST PROOFS]


(RAEF) Paper: RAEF-100942









Income and insurance effects of agricultural policy

Table 5. 1 . .. protection rate (ERP), real protection rates
(RRP) and insurance n.. (IE) with stochastic production,
USA, 1986-2000

Coefficient of relative Regime 1: Regime 2:
Risk aversion (R) 1986-1995 1996-2000
ERP11 = 0.96 ERP12= 0.36
ERP21= 2.71 ERP22= 1.67
1.0 RRP1 1.23 0.42
IE1 (%) 14 4
RRP2 3.34 1.82
IE2 (%) 17 6
2.0 RRP1 1.62 0.48
IE1 (%) 34 9
RRP2 4.22 1.98
IE2 (%) 41 12
3.0 RRP1 2.17 0.55
IE1 (%) 62 14
RRP2 5.50 2.15
IE2 (%) 75 18


Table 6. 17 .. protection rate (ERP), real protection rates
(RRP) and insurance n.. (IE) with stochastic production,
European union (15), 1986-2000

Coefficient of relative Regime 1: Regime 2:
Risk aversion (R) 1986-1992 1993-2000
ERP11 = 1.88 ERP12= 0.45
ERP1 = 2.17 ERP22= 1.78
1.0 RRP1 2.04 0.50
IE1 (%) 5 3
RRP2 2.35 1.90
IE2 (%) 6 4
2.0 RRP1 2.23 0.55
IE1 (%) 12 7
RRP2 2.56 2.02
IE, (%) 13 9
3.0 RRP1 2.42 0.61
IE1 (%) 18 11
RRP2 2.78 2.15
IE2 (%) 19 13


following the regime changes, the addition of all remaining
budgetary transfer payments dampened the decline attrib-
utable only to reduced incentive payments. For both
regimes, the producer incentive effect (ERPi) for the EU
was about double that of the USA However, when all
budgetary supports are included (ERP2i) both markets
maintained roughly equivalent income effects.

Insurance effect. The insurance effect is the percentage
change in the real rate of protection due to the stabilizing
effect of policy. For the USA (R = 2.0), the stabilizing ben-
efit of the pre-FAIR producer incentive policies contribu-
ted an additional (beyond the 'income effect') 34% increase
in producer incomes. This is an added benefit to producers.
The post-FAIR reforms lead to a substantial reduction of


the insurance benefit. With all policy support included, the
real rate of protection would be understated by 41% in
regime 1 but only 12% in regime 2. Risk averse producers
suffered real losses as a result of post-FAIR reforms.
For the EU (R = 2.0), the stabilizing benefit of the incen-
tive policies contributed an additional 12% increase in
expected utility of producer income during regime 1.
However, in the post-MacSharry period this insurance
effect fell to 7%. Producers place a relatively greater
value on the stabilizing characteristics of the pre- more
than the post-MacSharry policies. A similar relationship
is observed for all policy support.
In general, as regime 1 is moved from, which was char-
acterized by support based mostly upon price and output,
to regime 2, which was characterized by more decoupled
support, the relative insurance effects fall. This accords
with expectations in that policies in the latter regimes pro-
vided relatively more support through direct, decoupled
payments. This does not imply that overall support was
less but only that the proportion of that support delivered
through policies having substantial insurance effects was
smaller. Indeed, support levels could be large and the insur-
ance effects smaller as more of the support is delivered
through decoupled types of programmes. As expected,
insurance effects rise as risk aversion increases in that
risk-reducing programmes provided more benefits to highly
risk averse producers.
Also, for both markets, as a move is made from only
incentive supports to total budgetary support within
regime 2, an increase in the insurance effect is observed.
This is because the addition of large direct payments (e.g.,
market loss assistance payments in the USA) provided
an added measure of insurance. Despite policy reforms,
an important insurance benefit continues to be provided
to farmers.


V. CONCLUSIONS

The degree to which agricultural policy protects and sup-
ports agricultural producers is important. Traditional indi-
cators of protection refer only to the income effect. It is
argued that these measures by themselves are incomplete
since they do not account for the insurance effect resulting
from more stable incomes. The insurance effect is a mea-
sure of public sector insurance which is delivered through a
variety of government policies.
A way to incorporate both the income and insurance
effects in a comprehensive protection measure is proposed.
This indicator is termed the Real Rate of Protection. The
method is empirically illustrated with data from the US
and EU wheat markets for the period 1986-2000. Strong
evidence is found that the insurance effect is an important
component of protection, albeit a small one relative to the
income effect. This result accords with the conclusions of


[1-8] [Page No. 7] {Tandf}Raef/RAEF-100942.3d


[25.11.2003-8:53pm] [FIRST PROOFS]


(RAEF) Paper: RAEF-100942











Hennessy (1998), who found that decoupled payments had
relatively modest insurance effects. Relative to no policy
interventions, the average US farmer income was found
to increase 37% due to producer incentive support and
150% when all budgetary support is included. For the
EU, the corresponding increases were 77% and 162%,
respectively. While the income effect of total budgetary
support was similar for both the USA and EU, a greater
portion of total support was due to producer incentive
support in the EU than in the USA.
It was found that the insurance (stabilizing) effect of
producer incentive policies to be considerably greater for
the USA than the EU. When all market support incentives
and budgetary payments are included, both the USA and
EU policies provided insurance effects roughly 10-20%
greater than their income effects, with the larger insurance
benefit accruing to the more risk averse farmers.
Greater insight into the effects of coupled and decoupled
policies was obtained through an examination or policy
regime subperiods. For the USA, this was pre- and
post-FAIR Act regimes and for the EU pre- and post-
MacSharry reform periods. The post-FAIR and post-
MacSharry policies tended to be more decoupled than
their pre-regime change counterparts. It was found that
overall real rates of protection remained high following
the two reforms (although smaller than during the pre-
reform periods), while important insurance benefits con-
tinued to be delivered through the post-reform policies.
Decoupled support policies, however, tended to deliver rela-
tively smaller additional insurance benefits to producers
than coupled price and output policies alone.
Without accounting for the influence of policy on
income volatility, traditional measures of protection,
which rely on income levels only, may be misleading, at
least to the extent that policy benefits include a reduction
in the risk faced by farmers. In the case of US and EU
wheat, they will considerably underestimate the real rate
of policy protection.

ACKNOWLEDGEMENTS
This research was made possible with funding from
the Deutsche Forschungsgemeinschaft, The Ohio State
University, and the Austrian-American Fulbright Com-
mission. It was written while Thompson was a Fulbright
scholar at the University of Agricultural Sciences, Vienna,
Austria.


REFERENCES
2 European Commission (2001) Eurostat, Cronos Data Bank, EU.
Hennessy, D. A. (1998) The production effects of agricultural
income support policies under uncertainty, American
Journal of Agricultural Economics, 80, 46-57.
Hennessy, D., Babcock, B. A. and Hayes, D. (1997) The budget-
ary and producer welfare effects of revenue assurance,
American Journal of Agricultural Economics, 79, 1024-34.


S. R. Thompson et al.

Johnson, M. E. and Tenebein, A. (1981) A bivariate distribution
family with specified marginals, Journal American Statistical
Association, 76, 198-201.
Newbery, D. M. G. and Stiglitz, J. E. (1981) The Theory of
Commodity Price Stabilization.: A Study in the Economics of
Risk, Oxford University Press, Oxford.
OECD, Agricultural Policies in OECD Countries: Monitoring and
Evaluation, various issues, OECD, Paris.
Saha, A., Shumway, R. and Talpaz, H. (1994) Joint estimation of
risk preference structure using expo-power utility, American
Journal of Agricultural Economics, 76, 173-84.
Schmitz, M. (1997) CAP and food security, in Issues in
Agricultural Competitiveness: Markets and Policies (eds) R.
Rose, C. Tanner, and M. Bellamy, IAAE Occasional Paper
No. 7. International Association of Agricultural Economists,
Ashgate Publishing, Brookfield, Vermont.
Thompson, S. R., Gohout, W. and Herrmann, R. (2002a) CAP
reforms in the 1990s and their price and welfare implications:
the case of wheat, Journal Agricultural Economics, 53, 1-13.
Thompson, S. R., Herrmann, R. and Gohout, W. (2000)
Agricultural market liberalization and instability of domestic
agricultural markets: the case of the CAP, American Journal
of Agricultural Economics, 82, 718-26.
Thompson, S. R., Sul, D. and Bohl, M. T. (2002b) Spatial market
efficiency and policy regime changes: seemingly unrelated
error correction model estimation, American Journal of
Agricultural Economics, 84, 1042-53.
US Department of Agriculture, Commodity Costs and Returns at:
http://www.ers.usda.gov/data/costsandreturns/testpick.htm


APPENDIX


(Al)

(A2)


E[U(y)] = E[y][1 + 21o0.5R
S Ey]
(1 + cv)


Define the real rate of protection (RRP) as the percentage
change in the expected utility of income with protection
(i = 1) vis a vis without protection (i= 0).
Or


RRP


YI o


E[yl]/(1 + cv2y1)05
E[yo]/(1 +, ,,)0.5R
E[yl] (1 + cv2yo)05R
E[y0] (1 + cv2y)0.5R
K


(A3)


1 (A4)


1 (A5)


Now, let


E[y] E[yol
E[yo]


ERP


E[y1]
= (1 + ERP)
E[yo]
substituting into Equation (A5) gives,
RRP = (1 + EPR) K 1


(A6)


[1-8] [Page No. 8] {Tandf}Raef/RAEF-100942.3d


[25.11.2003-8:53pm] [FIRST PROOFS]


(RAEF) Paper: RAEF-100942









AUTHOR QUERIES

JOURNAL ID: RAEF-100942

QUERY QUERY
NUMBER
1 Both upper case and lower case CVs are used, which is correct
2 Provide publisher town
3 Provide page range




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs