The World’s Largest Open Access Agricultural & Applied Economics Digital Library

This document is discoverable and free to researchers across the

globe due to the work of AgEcon Search.

Help ensure our sustainability.

Give to AgEcon Search

AgEcon Search
http://ageconsearch.umn.edu
aesearch@umn.edu

Papers downloaded from AgEcon Search may be used for non-commercial purposes and personal study only.
No other use, including posting to another Internet site, is permitted without permission from the copyright
owner (not AgEcon Search), or as allowed under the provisions of Fair Use, U.S. Copyright Act, Title 17 U.S.C.
2015 Australian Agricultural and Resource Economics Society Conference

Contributed Paper

Steele Christian West (University of Western Australia)

The joint effects of technical efficiency and risk exposure on mixed enterprise farm output
variability in Western Australia

ABSTRACT

This study quantifies the importance of inefficiency and risk as sources of production variability in
Western Australian mixed crop - livestock broadacre farm businesses. Sources of farm level
observable heterogeneity are examined as determinants of inefficiency and risk through application
of Greene’s True Fixed Effects stochastic production framework in a Cobb-Douglas functional form.
Empirical Analysis is undertaken through a balanced panel of farm data from 274 operations
between 2002 and 2011. Results indicate output variability is mainly a consequence of risk as
opposed to technical inefficiency. Degree of production specialization, costs of finance, and capital
structure are shown to be significant to inefficiency. Production specialization, rainfall variability, and
capital structure are shown to be significant to and increase risk.

Keywords: risk exposure, technical inefficiency, stochastic frontier analysis, mixed enterprise farms

1. Introduction

For farm businesses, the technical inefficiency of farm production and the risks to which farm
production is exposed are jointly likely to influence farm output variability.

Only a few studies have chosen to examine technical efficiency of farm exposure and risk in
agricultural production. Tiedemann and Lataczs-Lohmann (2013) observed in their study of a small
sample organic and conventional farms in Germany find that variability of production risk has a
greater effect on output variability than technical inefficiency. Bokusheva and Hockman (2006) also
find that production risk has a greater relative effect on output variability than technical inefficiency.

A small number of authors have noted factors that effect production risk and technical inefficiency.
Villano and Fleming (2006) in their analysis of Filipino rice producers study the impact of a diverse
range of sociological, environmental and methodological factors on technical efficiency and
production uncertainty. Chang and Wen (2011) in a study of Taiwanese rice producers examine the
impact of off-farm income on technical efficiency and production risk and observe that farmers that
have off farm income were able to accommodate increased production risk, but not necessarily at
higher technical inefficiency. Jaenicke, Frechette, and Larson (2003) investigate the effect of input
use on inefficiency and production risk for cotton production in West Tennessee.

In Western Australia, Mugera and Nyambane (2014) find that for broadacre farms technical
efficiency is positively influenced by short term debt, tax liability and capital investment, whilst
negatively influenced by off-farm revenue generating activities.

In Australian agriculture more broadly, there are several studies that examine technical efficiency in
farm production (Battese, Coelli 1995; Doucouliagos, Hone 2000; Fraser, Hone 2001; Fraser, Horrace
2003; Kompas, Che 2006) or examine changes in total factor productivity and its components
(Nossal, Sheng, Zhao, Gunasekara 2009; Tozer, Villano 2013; Sheng, Zhao, Nossal, Zhang 2014; Islam,
Xayavong, Kingwell 2014) . Climate variability is a key feature of Western Australian agriculture
(CSIRO & Bureau of Meteorology 2007; Hennessy et al 2008) and adverse risk from climate change
presents substantial risk for farmers in southern Australia (Garnaut 2010; Asseng, Pannell 2012),
which indicates that considerable merit exists for the joint study of production risk and technical
inefficiency.

The present study proposes to determine the contributions of risk and technical inefficiency to
output variability for mixed crop-livestock farms in south west Western Australia through the
application of a ‘true effects’ stochastic frontier analysis. The study identifies sources of observable
heterogeneity amongst these farms that significantly affect risk.

The farms in the present study are broadacre dryland operations that receive low levels of
government assistance and subsidization relative to farm operators in several other developed
countries.

The paper is organized as follows: section 2 provides an overview of the prior studies of technical
efficiency and risk; section 3 details the analytical framework and the data used; section 4 presents
the empirical findings and section 5 states the study conclusion and implications.

2. Technical efficiency and production risk in farm business

Technical Efficiency represents the effectiveness with which a given set of inputs is used to produce
an output (Farrell 1957). Many sources of observable heterogeneity between farms globally have
been shown in prior studies to significant affect the farm’s technical efficiency.
Studies of capital structure and technical efficiency (Lambert, Bayda 2005; Emvalomatis, Oude
Lansik, Stefanou 2008) have provided divergent results. Some results provide support for both
Agency theory (Jensen, Meckling 1976) and free cash flow theory (Jensen 1984). Free cash flow
theory asserts that higher debt usage will increase technical efficiency, since management will need
to exercise increased vigilance to avoid the negative consequences of failure to service their
obligations. Conversely, agency theory proposes that debt and technical efficiency would be
inversely related, as a consequence of the difficulty associated to lenders being able to monitor
borrowers and hence imposing higher costs of credit.

Analysis of the impact of credit constraints on technical efficiency in agriculture (Blancard,

Boussemart, Briec, Kerstens 2006; Davidova, Latruffe 2007) suggests the possible presence of both
agency theory and signalling theory (Ross 1977, Hubbard 1998), where the preferences of lenders
affect farm investment capacity and hence technical efficiency. Increased investment, for example,
has been observed to increase technical efficiency (Doucogliagos, Hone 2000; Kumbhakar,
Bokusheva 2009).

Production specialization (Featherstone, Langemeier, Ismet 1997; Bokusheva, Hockman, Kumbhakar

2012) is an indicator of resource allocation and input use, and is also a likely influence on technical
efficiency. Production specialization should allow farmers to concentrate on specific production
processes and increase technical efficiency. Increased education and experience (Dhungana, Nuthall,
and Nartea 2004) in theory should translate to increased skill and knowledge, which also should
promote increased technical efficiency.

The significance of the effect of farm size (Byrnes, Färe, Grosskopf, Kraft 1987; Hallam, Machado
1996; Mugera, Langemeier 2011), subsidisation (Serra, Zilberman, Gil 2008), and technology choice
(Kompas, Nhe Che 2006; Mayan, Balagtas, Alexander 2010) on technical efficiency has also been
addressed in prior studies.

In Western Australia, Mugera and Nyambane (2014) observed short term debt use, increased tax
liabilities (a consequence of increased profitability) and capital investment were important in raising
technical efficiency, a finding which is consistent Sheng, Zhao, Nossal, Zhang’s (2014) study of how
new production technology can increase production efficiency.

Chavas (2008) identified two primary sources of risk in price uncertainty (i.e. market prices for inputs
and outputs) and production uncertainty (such as industrial action, climate, and technological
change). Uncertainty of demand and the irreversibility of investment decisions have been shown in
the context of farms in the south east of the United States to influence investment decisions (Isik,
Coble, Hudson, House 2003) and land development decisions in the Kyrgyz Republic (Savastano,
Scandizzo 2009). Output price volatility can affect the global crop acreage (Haile, Kalkuhl, von Braun
2013), while input price stability has promoted increased adoption of new technology (Schonegold,
Sunding 2014).

Technological progress has been shown by Kim and Chavas (2003) to reduce farmer’s risk exposure
and downside risk. Regulatory policy has been shown to influence farmer risk perception. For
example, Koundouri, Laukkanen, Myyra, Nauges (2009) examine the increase in non-random income
components of Finnish farmers following Finland’s accession into the European Union. They found
that the EU’s decoupling policies affected farmer’s input use and crop use through adjustment of
farmer’s risk attitudes. Increased environmental uncertainty has been shown to induce an increase
in production diversification by farmers to mitigate such risks (Baumgärtner, Quaas 2009).
Production specialization would be anticipated to increase production risk, based on the application
of portfolio theory (Markowitz 1952).

3. Methodology and Data

3.1 Theoretical Modelling

This study uses stochastic frontier analysis (‘SFA’) to determine the impact of observable farm level
heterogeneity on technical efficiency and risk in Western Australian farm businesses. SFA is a
parametric method that invokes assumptions about parameters’ random errors.

SFA was first proposed as an extension of prior deterministic studies by Aigner, Lovell and Schmidt
(1977) who applied a half normal distribution of the error term. Independently, Meeusen and Van
den Broeck (1977) applied an exponential distribution. The adopted functional form of the SFA
model used in thre present study follows that proposed by Aigner et al:

yit  f ( xit , zi )  vit  uit   ' xit   ' zi  vit  uit

i  1,..., N , t  1,..., T

vit ~ N [0,  v2 ]

uit  Uit , where U it ~ N [0,  u2 ]  vit

In the above stated function, yit represents output, xit represents a vector of inputs or input prices, zi
is a vector of firm specific characteristics, vit is a random error associated to factors beyond the
production entity’s control (weather, political or economic shocks etc), uit represents of inefficiency,
i represents an individual producer and t represents an individual production period.
 3.2 Empirical Modelling

Construction of the study variables is outlined in Appendix 1. A Box-Cox transformation (Box, Cox
1964) is applied to generate a functional form:

y 1
y 

The Box-Cox transformation tests four models:

(i) Theta- independent and dependent variables subject to a separate transformation:

yi  1 x1j  2 x2 j  ...  2 x2 j   j

(ii) Lambda- independent and dependent variable subject to a common transformation:

yi  1 x1j  2 x2 j  ...  2 x2 j   j

(iii) Right Hand Side- dependent variable only subject to a transformation:

yi  1 x1 j  2 x2 j  ...  2 x2 j   j

(iv) Left Hand Side- independent variable only subject to transformation:

yi  1 x1j  2 x2 j  ...  2 x2 j   j

This study directs specific attention to the test of three common functional forms in application of
the Box Cost test:

Linear: y   y  1 if λ=1

Log specification: y   ln( y) if λ=0

1
Multiplicative inverse: y   1  if λ=-1
y

Post specification of functional form, a Hausman Test (Hausman 1978) was utilised to differentiate
between whether the panel data was subject to fixed and random effects. A Hausman test has a null
hypothesis (H0) that the random effects estimator (b1) is preferred as it consistent and efficient;
under the alternative hypothesis (HA), the fixed effects (b0) estimator is preferred since it is at least
consistent. In consideration of a standard linear model y=bX+e, the Wu-Hausman Test Statistic is:

H  (b1  b0 )' (Var (b0 )  Var (b1 ))† (b1  b0 ) ,

Where † indicates a Moore-Penrose pseudo inverse1.

The Hausman test indicates that a Fixed Effects model is preferred (refer ‘4.Results’). A fixed effects
SFA estimator (see Schmidt, Sickle 1984; Cornwell, Schmidt, Sickles 1990; Kumbhakar 1990; Lee,
Schmidt 1993) is free of distributional assumptions and requires only the statement of the
conditional mean; it also allows for correlation between effects and time varying regressors. These
benefits, however, are somewhat negated in the above cited estimators by the loss of the individual
identity in the conventional fixed effects formulation as stated below:

yit     ' xit  Sui  vit

 i   ' xit  vit ,

Where i    Sui

The loss of this identity is because the effects are only measured relative to the ‘best’ (most
efficient) within the sample.

Estimation of the stochastic frontier model in this study is undertaken through application of an
extended ‘true’ fixed effects (‘TFE’) model as proposed by Greene (2005, 2005a), which addresses
the loss of individual identity. This model provides an important advancement of prior fixed effects
formulations that are derivative of the Schmidt and Sickles (1984) formulation ( yit  i   ' xit  vit )
in that time variant inefficiency, uit , is separated from i , a group specific constant. The
problematic non-consideration of time variant inefficiency and the preclusion of covariates that do
not vary through time are also problems that this approach removes (Greene 2005a). Furthermore,
heterogeneity may be correlated with group variables under the TFE approach. Consistent with the
presence of heteroscedasticity in both error terms vit and uit , the TFE model is stated as:

yit   i   ' xit  vit  uit

uit ~ N  (0,  uit
2
)
vit ~ N (0,  vit2 )
 uit2  g ( zit ;  )
 vit2  h( zit ;  )

Consistent with the specification of Greene (2005a), the log likelihood function is estimated as a
fixed effects model:

1
Moore Penrose pseudoinverse: M(m,n;K), where m,n is a vector of m x n matrices and K is representative of R
or C. For A  M(m,n;K) a pseudo inverse of A is matrix A+  M(m,n;K) s.t.:
i)AA+A=A
ii) A+AA+=A+
iii)(AA+)*=AA+
iv)(A+A)*=A+A
 N T  1   y   it   ' xit    yit   i   ' xit 
LogL   log      it     ,
i 1 i 1  (0)        

where is the  standard normal Cumulative Density Function and  is the standard normal density.

Post maximization of the Log Likelihood function, the JLMS estimator (Jondrow, Materov, Lovell,
Schmidt 1982) is used to estimate uit given:

   ( it ) 
 uit  it   2 
  it 
1   1   ( it ) 
 it  vit  uit  yit     xit
1
  [ v2   u2 ] 2

 u
v

 it   it


Where  (it ) is the standard normal density and (it ) is the cumulative density function
evaluated at it (Greene 2005).

Estimation of technical efficiency allows for the calculation of the proportions of output variability
attributable to inefficiency and risk. Subject to the assumption of a half normal distribution for the
inefficiency term, the calculation proposed by Kumbhakar and Lovell (2003) was utilized where π is
the net profit of the operation:

 2 2
 2   v2  Varu   v2    u
  

3.3 Study region and farm data

The farm data used in the analysis covers the period from 2002 to 2011. 274 farms located in south
west Western Australia who engaged one of three major agricultural consultancies (PlanFarm, Evans
& Grieve, and Farmanco) collected annual data on farm operations and finances. Only farms that
provided data for all ten seasons were included in this data set. This induces some potential bias in
the failure to capture the entry and exit of farm businesses (see Foster, Haltiwanger, Syverson 2008).

South west Western Australia is characterized by large scale broadacre dryland farms that operate
crop, mixed, and/or livestock production subject to a Mediterranean climate. The primary crops are
wheat, barley, canola, lupins and oats. Farms produce one dryland crop per annum. Sheep account
for the majority of livestock held on these farms. The smallest property surveyed between 2002 and
2011 was 365 Hectares, the largest was 16,988 Hectares.
A broad range of information was recorded in the survey including items such as annual rainfall, land
size and allocation, labour use, crop production values and quantities, variable and fixed cost
expenditure values, financial particulars inclusive of farm income, asset and liability measurements,
farm owner characteristics inclusive of educational attainment, age range, and family structure, as
well as producer and consumer price indexes.

3.4 Index of variables

As per the requirements of the preferred methods, an output variable, input variables and variables
that account for observable heterogeneity were constructed from the data set. Table 1 provides a
summary of the variables used in this analysis; for an explanation as to the construction of the
variables, refer to Appendix 1.

Table 1. Variable Summary

Measure µ σ 95% Conf. Interval

Dependent Variable
Output (y) 8514.378 139.0449 8241.735 8787.021
Input Variables
Labour (x1) 164.6722 8.796342 147.424 181.9203
Crop inputs (x2) 3261.089 48.33898 3166.305 3355.874
Operational costs (x3) 1054.119 14.75511 1025.187 1083.052
Livestock production inputs (x4) 1865.226 35.87414 1794.883 1935.569
Growing season rainfall (x5) 242.9417 1.796394 239.4192 246.4641
Observable Heterogeneity: Inefficiency
Production specialization (z1u) -.444960 .0064421 -.457592 -.432328
Cost of finance (z2u) -3.10806 .0229454 -3.15305 -3.06306
Capital structure (z3u) -1.57787 .0195554 -1.61622 -1.53953
Experience (z4u) 2.822458 .0142255 2.794562 2.850353
Education (z5u) 1.448458 .0160055 1.417071 1.479845
Observable Heterogeneity: Uncertainty
Production specialization (z1v) -.444960 .0064421 -.457592 -.432328
Capital Structure (z2v) -1.57787 .0195554 -1.61622 -1.53953
Price variability index (z3v) .3226207 .0012094 .3202493 .3249921
Rainfall variability index (z4v) .3413043 .0029839 .3354534 .3471552
Regulatory change- Wheat Export .6 .0093437 .5816786 .6183214
Marketing Act 2008 (z5v)

3.5 Model Estimation

Following the program method set forth in Belotti, Daidone, Ilardi and Atella (2012), the panel data
set was analysed using STATA.

4. Results

The initial test undertaken was for the model specification as per the BoxCox test. Investigation of
the theta, lambda, right hand side and left hand side transformations yielded only one common
functional form as nominated in Section 3.2 that was not strongly rejected. This was the lambda
restriction whereupon both the dependent and independent variables were transformed subject to
a lambda equal to zero.

Table 2. BoxCox Test Results

Test Restricted LR statistic P-Value

H0 Log Likelihood χ2 Pr(>χ2 )
λ= -1 -27491.659 10624.29 0.000
λ=0 -22179.856 0.69 0.407
λ=1 -23415.786 2472.55 0.000

The test result in Table 2 indicates that a logrithmic transformation cannot be strongly rejected.

Table 3. Hausman Test

(b) (B) (b-B) √(diag(V_b-V_B))

Fixed Random Difference Standard Error
Ln(x1) .13058 .1997125 -.0691326 .0106289
Ln(x2) .0506916 .052157 -.0014654 .0033765
Ln(x3) -.021635 -.028098 .0064636 .0042172
Ln(x4) .009864 .0087022 .0011618 .0021929
Ln(x5) .6499038 .6346067 .0152972 .0073728

Test: H0- difference in coefficients not systematic

χ2 (5) = (b-B)'[(V_b-V_B) -1](b-B)

= 52.19

Pr(>χ2 )= 0.0000

The Hausman test confirms the rejection of the null hypothesis that individual-level effects are
adequately modelled by a random-effects model. Henceforth a fixed effects model is instituted.

In accordance with the results of the Hausman test, Greene’s true fixed effects model is applied to a
Cobb Douglas function transformed as per the Box Cox test results.

Table 4. Stochastic frontier analysis- production inputs

Frontier Coef. Std. Err. z P>z 95% C.I.

Ln(x1) .1139937 .0236324 4.82 0.000*** .067675 .1603123
Ln(x2) .0408997 .020476 2.00 0.046** .0007675 .0810318
Ln(x3) .003998 .0243186 0.16 0.869 -.043665 .0516616
Ln(x4) .0127434 .010143 1.26 0.209 -.007136 .0326233
Ln(x5) .4932979 .0286455 17.22 0.000*** .4371538 .549442
*= 10% significance, **= 5% significance, ***= 1% significance
As per the functional form specified, the coefficients estimated represent the output elasticities of
each of inputs, with rainfall shown to have the greatest effect followed by labour. Both are
significant at a one percent level.

The inefficiency coefficients estimated by the true fixed effects model are detailed in Table 5A that
shows that production specialization, costs of finance and financial risk aversion are all significant at
a 1% level. Increased crop specialization is shown to reduce inefficiency, while higher costs of
finance and debt use are shown to increase inefficiency. Education and Experience were both shown
to reduce inefficiency, though neither was significant.

Table 5A. Analysis of sources of farm level observable heterogeneity on technical inefficiency

σu Coef. Std. Err. Z P>z 95% C.I.

z1u -2.50477 .2401946 -10.43 0.000*** -2.97554 -2.03399
z2u .3437249 .0765304 4.49 0.000*** .1937281 .4937216
z3u -.288143 .086803 -3.32 0.001** -.458274 -.118012
z4u -.028306 .129989 -0.22 0.828 -.283080 .2264674
z5u -.059130 .1265258 -0.47 0.640 -.307116 .1888555
constant -2.61676 .5221552 -5.01 0.000*** -3.64016 -1.59335

The reduction in inefficiency associated with increased crop specialization is in accord with the
findings of Bokusheva, Hockman, and Kumbhakar (2012) in their study of Russian agriculture from
1999 to 2009. Increased crop specialization may allow for increased mechanization, which promotes
increased technical efficiency. The finding that increased debt is negative and significant to technical
inefficiency lends support to free cash flow theory and indicates that farmers become more diligent
when faced with the heightened penalty of default. The positive and significant impact of borrowing
costs (capital constraints) on technical inefficiency is further in accord with agency theory and is
consistent with the findings of Blancard, Boussemart, Briec, and Kerstens (2006) in their study of
capital and expenditure constraints on farms in Nord-pas-de-Calais, France

The negative impact of age on technical inefficiency indicates that increased experience promotes
technical efficiency. The negative relationship between education and technical efficiency indicates
that farmers with higher educational attainment are more technically efficient.

Estimation of risk in response to sources of observable heterogeneity indicates that production

specialization, price variability and rainfall variability are positive and significant at a 1% level
regarding production risk. As crop production as a percentage of total production increases, so does
production risk. This finding is consistent with theoretical expectation as specified by portfolio
theory (Markowitz 1952). Increased risk as a consequence of a higher debt to equity ratio is also in
direct alignment with theoretical expectation. The introduction of the Wheat Export Market Act in
2008 that deregulated wheat export marketing in Australia is not significant in affecting risk.

Table 5B. Analysis of sources of farm level observable heterogeneity on risk

σv Coef. Std. Err. Z P>z 95% C.I.

z1v 2.03733 .3063006 6.65 0.000*** 1.436992 2.637668
 z2v .3612944 .0872042 4.14 0.000*** .1903774 .5322114
z3v .9138366 .6659282 1.37 0.170 -.391358 2.219032
z4v 1.588657 .4231207 3.75 0.000*** .7593559 2.417959
z5v .0821041 .119164 0.69 0.491 -.151453 .3156613
constant -2.18009 .2945215 -7.40 0.000*** -2.75734 -1.60284

The next stage of the estimation is the estimation of technical inefficiency, u ; this is done through
application of the JLMS estimator (refer section 3.2). Post estimation of u , the variance of the
inefficiency term,  u2 , and the variance of output,  y , are used to calculate risk variance,  v2 , as
2

per the method set forth by Khumbakar and Lovell (2003). In application of this method, variability
of risk (  v = 0.6669) is shown to have a substantially greater impact on output variability (  y =

0.7185) than variability of technical inefficiency (  u = 0.2675).

In comparison of the coefficients obtained from the inefficiency and risk variable analysis, it is
observed that production specialization and capital structure have a significant and positive effect
on risk while having a significant and negative impact on technical inefficiency. As output variability
is more strongly influenced by risk variability than technical inefficiency variability, this supports the
premise that farmers should seek to prioritise actions that reduce risk variability.

Reduction in cost of capital, positive and significant to technical inefficiency, may represent the best
means to address output variability for Western Australian farmers. Reduction in the cost of capital
would increase the accessibility of technology to diversify production and allow for investment in
technologies that could reduce technical inefficiency. Decreased borrowing charges would also
lower total liabilities for a fixed amount, or alternately allow farmers to borrow more money for
equal repayments.

Reductions in borrowing costs for farmers could be promoted through initiatives that decrease
asymmetry between the information available to borrowers and lenders in consonance with agency
theory; this would be to the mutual benefit of farmers and lenders as it would reduce the business
risk of both parties.

5. Conclusion

This article is the first in the context of Australian agriculture that seeks to quantify risk and technical
inefficiency conjunctively to determine their relative impact on output variability. The data
considered is a balanced panel of 274 farms for the sample period of 2002 to 2011. A stochastic
frontier analysis is undertaken subject to a true fixed effects specification as defined by Greene
(2005) that allows for the separate identification of time variant inefficiency and risk. Sources of
observable heterogeneity amongst farms are examined as exogenous variables in the variance
functions of the time variant inefficiency and risk to determine their significance to these conditions.
Through the application of the JLMS estimator and the output variability decomposition of
Kumbhakar and Lovell (2003), the standard deviations of inefficiency and uncertainty are calculated
to examine their relative effect on the variability of output.
The following conclusions may be drawn from this study. First, the study finds that variability in risk
has a greater effect on output variability in the context of mixed crop – livestock operations in
Western Australia than variability in technical efficiency does. Second, the study finds that
production diversification and capital structure are important factors in determination of both
technical efficiency and uncertainty at the farm level; increased specialization and debt use is
associated with a reduction in technical inefficiency while both increase uncertainty. Increased risk
as a consequence of increased volatility in rainfall and output prices is directly concordant with
theoretical expectation. The significance of higher interest costs to increased technical inefficiency
indicate the perception of farm business quality in lending is a significant driver of technical
efficiency for mixed output farm businesses in Western Australia.

These findings suggest that farmers in Western Australia will substantially benefit from policy that
promotes the mitigation of capital costs through the promotion of information symmetry and
transfer mechanisms. Policy that better educates farmers in the presentation of information to
financial lenders and financial management may assist in this regard. As variability of production risk
is more significant to output variability than the variability of technical inefficiency, initiatives that
promote production diversification could also offer positive benefits and security for farmers.
APPENDIX 1

Table A1. Variable Construction

Measure Description
Dependent Variable
Output (y) The total revenue of farm operations that have been normalized
through application of an overall consumer price index figures with a
2002 base year.
Production Inputs
Labour (x1) The aggregate of both casual labour and permanent labour used on a
farm; measured in weeks.
Crop inputs (x2) This variable was constructed as a three step process. First, the
expenditure on fertilizer, chemicals, seeds and fuel were normalized
over their respective consumer price index figures with 2002 assumed
as a base year. This is done since actual price data is not available.
Operational costs (x3) This variable was constructed as per the Crop Input variable except with
the original input expenditures being contract services (exclusive of
labour), administration, and repairs and maintenance expenditure.
Livestock production inputs (x4) Again this variable was constructed through the process of
normalization of individual component’s expenditure levels with the
subsequent aggregation of these inputs. The inputs used were livestock
purchased and expenditure on livestock production.
Growing season rainfall (x5) This is the rainfall recorded between April and November, which is the
growing season in South West region of Western Australia.
Observable Heterogeneity: Inefficiency
Production specialization (z1u) Farm specialization was represented by the natural log function of the
land area under production used for crop production divided by the
total land area under production.
Cost of finance (z2u) The natural log of the ratio of interest expenses as a percentage of total
liabilities is used to highlight heterogeneity in the cost of finance for
farms.
Capital Structure (z3u) This was represented through the natural log of the ratio of total
liabilities to total equity, i.e. the farm’s capital structure. An increase in
this ratio is indicative of reduced risk aversion.
Experience (z4u) This was represented by the farm operator’s age. In the data surveyed,
only banded data was provided with classification ranges of 30-45, 45-
60, 60-70, and 70+. The variable was constructed by through
application of the encode function is Stata to convert the survey results
to a format conducive for statistical analysis.
Education (z5u) This has been represented by farm operator education. The data
surveyed provides three banded results: Secondary, Tertiary Technical
and Tertiary University. These were converted for statistical analysis by
the application of the encode function in Stata.
Observable Heterogeneity: Uncertainty
Production specialization (z1v) See z1u

Capital Structure (z2v) See z3u

Price variability index (z3v) The natural log of the ratio of crop values over aggregate crop
production was calculated for each year for each farm. The standard
deviation of this function was calculated based on the ten years
available for each farm on a per farm basis. 70% of farmers in the study
data set were 45 years of age or older; as a result their decision making
can be assumed to be based on information accrued over a longer
period. Further it may be assumed that a farmer who experienced
increased price variability in the period of 2002 to 2011 could be
anticipated to have experienced increased price variability in prior
periods.
Rainfall variability index (z4v) The natural log of growing season rainfall is calculated for each farm.
The standard deviation of this function was calculated based on the ten
years available for each farm on a per farm basis. 70% of farmers in the
study data set were 45 years of age or older; as a result their decision
making can be assumed to be based on information accrued over a
longer period. Further it may be assumed that a farm who experienced
increased rainfall variability in the period of 2002 to 2011 could be
anticipated to have experienced increased price variability in prior
periods.
 Regulatory change (z5v) A dummy variable was constructed to account for the introduction of
the federal wheat export marketing act of 2008, with a score of ‘1’
representative of years prior to 2008 and ‘0’ representative of years
after

References

Aigner D J, Lovell C A K., Schmidt P (1977) “Formulation and estimation of stochastic frontier
production function models” Journal of Econometrics 6 (1), 21-37

Baumgärtner S, Quaas M F “Managing increasing environmental risks through agrobiodiversity and

agrienvironmental policies.” Agricultural Economics 41, 483-496

Belotti F, Daidone S, Ilardi G, Atella V (2012) “Stochastic frontier analysis using Stata.” CEIS Tor
Vergata- Research Paper Series 10 (12), No. 251

Blancard S, Boussemart J-P, Briec W, Kerstens K (2006) “Short- and Long-Run Credit Constraints in
French Agriculture: A directional distance function framework using expenditure-constrained Profit
Functions.” American Journal of Agricultural Economics 88 (2), 351-364

Bokusheva R, Hochmann H (2006), “Production risk and technical inefficiency in Russian agriculture”.
European Review of Agricultural Economics 33 (1), 93-118

Bokusheva R, Hochmann H, Kumbhakar S C (2012), “Dynamics of productivity and technical

efficiency in Russian Agriculture”. European Review of Agricultural Economics 39 (4), 611-637

Box G E P, Cox D R (1964) “An analysis of transformations.” Journal of the Royal Statistical Society,
Series B, 26, 211–252

Chao H, Wen F (2011) “Off-farm work, technical efficiency, and rice production risk in Taiwan.”
Agricultural Economics 42, 269-278

Chavas J P (2008) “A cost approach to economic analysis under state-contingent production

uncertainty.” American Journal of Agricultural Economics 90, 435-446

Cornwell C, Schmidt P, Sickles R (1990) “Production Frontiers with Cross-Sectional and Time-Series
Variation in Efficiency Levels.” Journal of Econometrics 46 (1), 185-200

CSIRO and Bureau of Meterology (2007) “Climate change in Australia”, Technical Report, CSIRO
Publishing, Melbourne

Davidova S, Latruffe L (2007) “Relationships between Technical Efficiency and Financial Management
for Czech Republic Farms.” Journal of Agricultural Economics 58 (2), 269-288

Dhungana B R, Nuthall P L, Nartea G V (2004) “Measuring the economic inefficiency of Nepalese rice
farms using data envelopment analysis. The Australian Journal of Agricultural and Resource
Economics 48 (2), 347-369

Doucouliagos H, Hone P (2000) “The efficiency of the Australian dairy processing industry.” The
Australian Journal of Agricultural and Resource Economics 44 (3), 423-438
Emvalomatis G, Oude Lansik A, Stefanou S E (2008) “An Examination of the Relationship between
Subsidies on Production and Technical Efficiency in Agriculture: The Case of Cotton Producers in
Greece.” 107th EAAE Seminar "Modelling of Agricultural and Rural Development Policies", Sevilla,
Spain, January 29th -February 1st, 2008

Farrell M J (1957) “The measurement of productive efficiency.” Journal of Royal Statistical Society
120, 253-290

Foster L, Haltiwanger J, Syverson J (2008) “Reallocation, Firm Turnover and Efficiency: Selection on
Productivity or Profitability?” American Economic Review 98 (1), 394-425

Fraser I M, Horrace W C (2003) “Technical Efficiency of Australian Wool Production: Point and
Confidence Interval Estimates.” Journal of Productivity Analysis 20, 169-190

Gardebroek C, Chavez M D, Oude Lansink A (2010) “Analysing Production Technology and Risk in
Organic and Conventional Dutch Arable Farming using Panel Data.” Journal of Agricultural Economics
61(1), 60-75

Garnaut R (2010) “Climate change and the Australian agricultural and resources industries.” The
Australian Journal of Agricultural and resource Economics 54, 9-25

Greene W (2005) “Fixed and Random Effects in Stochastic Frontier Models.” Journal of Productivity
Analysis 23, 7-32

Greene W (2005a) “Reconsidering Heterogeneity in panel data estimators of the stochastic frontier
model.” Journal of Econometrics 126, 269-303

Gunasekera D, Kim Y, Tulloh C, Ford M (2007), “Climate change impacts on Australian agriculture”.
Australian Commodities 14, 657-676

Haile M G, Kalkuhl M, von Braun J (2014) “Inter- and intra-seasonal crop acreage response to
international food prices and implications of volatility.” Agricultural Economics 45, 1-8

Hausman J A (1978) “Specification tests in econometrics.” Econometrica 46, 1251–1271

Hennessy K, Fawcett R, Kirono D, Mpelsoka F, Jones D, Bathols J, Whetton P, Stafford-Smith M,

Howden M, Mitchell C, Plummer N (2008) “An assessment of the impact of climate change on the
nature and frequency of exceptional climatic events”. A consultancy report by CSIRO and the
Australian Bureau of Meteorology for the Australian Bureau of Rural Sciences, pp. 33,
www.bom.gov.au/climate.droughtec/

Hubbard R (1998) “Capital Market Imperfections and Investment.” Journal of Economic Literature 36
(1), 193-225

Isik M, Coble K H, Hudson D, House L O (2003) “A model of entry-exit decisions and capacity choice
under demand uncertainty” Agricultural Economics 28, 215-224

Jaenicke E C, Frechette D L, Larson J A (2003) “Estimating Production Risk and Inefficiency

Simultaneously: An Application to Cotton Cropping Systems.” Journal of Agricultural and Resource
Economics 28 (3), 540-577
Jensen M C (1984) “Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers.” The
American Economic Review 76 (2), 323-329

Jensen M C, Meckling W H (1976) “Theory of the firm: Managerial behaviour, agency costs and
ownership structure.” Journal of Financial Economics 3 (4), 305-360

Jondrow J, Materov I, Lovell K, Schmidt P (1982) “On the estimation of technical efficiency in the
stochastic production frontier production model.” Journal of Econometrics 19, 233-238

Just R E, Pope R D (1978) “Production Function Estimation and Related Risk Considerations.” Journal
of Econometrics 7(1), 67-86

Kim K, Chavas J (2003) “Technological change and risk management: an application to the economics
of corn production.” Agricultural Economics 29, 125-142

Kingwell R, Farré I (2009) “Climate change impacts on investment in crop sowing machinery.” The
Australian Journal of Agricultural and Resource Economics 53, 265-284

Kimura S, Antón J (2011) “Risk Management in Agriculture in Australia”. OECD Food Agriculture, and
Fisheries Working Papers No. 39, OECD Publishing

Kompas T, Nhu Che T (2006) “Technology choice and efficiency on Australian dairy farms.” The
Australian Journal of Agricultural and Resource Economics 50, 65-83

Koundouri P, Laukkanen M, Myyrä S (2009) “The effects of EU agricultural policy changes on farmers’
risk attitudes.” European Review of Agricultural Economics 36 (1), 53-77

Kumbhakar S C (1990) “Production Frontiers, Panel Data, and Time-Varying Technical Inefficiency.”
Journal of Econometrics 46, 201-211

Kumbhakar S C (2002) “Specification and Estimation of Production Risk, Risk Preferences and
Technical Efficiency.” American Journal of Agricultural Economics 84 (1), 8-22

Kumbhakas S C, Lovell C A K (2003) “Stochastic Frontier Analysis.” Cambridge University Press

Kumbhakar S C, Bokusheva R (2009) “Modelling farm production decisions under an expenditure

constraint.” European Review of Agricultural Economics 36 (3), 343-367

Kumbhakar S C, Lien G, Hardaker J B (2014) “Technical efficiency in competing panel data models: a
study of Norwegian grain farming.” Journal of Productivity Analysis 41, 321-337

Lee Y, Schmidt P (1993) “A Production Frontier Model with Flexible Temporal Variation in Technical
Inefficiency.” In: Fried H, Lovell K, The Measurement of Productive Efficiency: Techniques and
Applications, New York: Oxford University Press

Loch A, Hatt M, Mamum E, Xu J, Bruce S, Heyhoe E, Nicholson M, Ritman K (2012) “Farm risk
management in a changing climate”. ABARES conference paper 12.5, Canberra, March 2012

Markowitz H (1952) “Portfolio Selection.” The Journal of Finance 7 (1), 77-91

Mayan C D, Balagtas J V, Alexander C E (2010) “Technology Adoption and Technical Efficiency:
Organic and Conventional Dairy Farms in the United States.” American Journal of Agricultural
Economics 92 (1), 181-195

Meeusen W, Van den Broeck J (1977) “Efficiency Estimation from Cobb-Douglas Production
Functions with Composed Error.” International Economic Review 18 (2), 435–444

Mugera A W, Langemeier M R (2011) “Does Farm Size and Specialization Matter for Productive
Efficiency? Results for Kansas.” Journal of Agricultural and Applied Economics 43 (4), 515-528

Mugera A W, Nyambane G G (2014) “Impact of debt structure on production efficiency and financial
performance on Broadacre farms in Western Australia.” Australian Journal of Agricultural and
Resource Economics 56, 1-17

Pannell D J, Malcolm B, Kingwell R S (2000) “Are we risking too much? Perspectives on risk in farm
modelling.” Agricultural Economics 23, 69-78

Pitt M, Lee L (1981) “The Measurement and Sources of Technical Inefficiency in Indonesian Weaving
Industry.” Journal of Development Economics 9, 43-64

Ross S A (1977), “The Determination of Financial Structure: The Incentive-Signalling Approach.” The
Bell Journal of Economics 8 (1), 23-40

Savastano S, Scandizzo P L (2009) “Optimal farm size in an uncertain land market: the case of Kyrgyz
Republic.” Agricultural Economics 40 (2009), 745-758

Schmidt P, Sickles R (1984) “Production Frontiers with Panel Data.” Journal of Business and Economic
Statistics 2 (4), 367-374

Schoengold K, Sunding D L (2014) “The impact of water price uncertainty of the adoption of
precision irrigation systems.” Agricultural Economics 45, 1-15

Serra T, Zilberman D, Gil J M (2008) “Farms’ technical inefficiencies in the presence of government
programs.” The Australian Journal of Agricultural and Resource Economics 52, 57-76

Serra T, Stefanou S, Oude Lansink A (2010) “A dynamic dual model under state-contingent
production uncertainty.” European Review of Agricultural Economics 37 (3), 293-312

Sheng Y, Mullen J D, Zhao S (2011) “A turning point in agricultural productivity: consideration of the
causes”. ABARES research report 11.4 for the Grains Research and Research and Development
Corporation, Canberra, May 2011

Sheng Y, Zhao S, Nossal K, Zhang D (2014) “Productivity and farm size in Australian agriculture:
reinvestigating the returns to scale.” Australian Journal of Agricultural and Resource Economics 58,
1-23

TiedemannT, Latacz-Lohmann U (2013) “Production Risk and Technical Efficiency in Organic and
Conventional Agriculture – The Case of Arable Farms in Germany.” Journal of Agricultural Economics
64 (1), 73-96
Van Biesebroeck J (2007) “Robustness of Productivity Estimates.” The Journal of Industrial Economics
LV, 529-569

Villano R, Fleming E (2006) “Technical Inefficiency and Production Risk in Rice Farming: Evidence
from Central Luzon Philippines.” Asian Economic Journal 20 (1), 29-46