Do Macroeconomic Conditions Matter for Agriculture?

The Indian Experience

by
Shashanka Bhide, B.P. Vani and Meenakshi Rajeev *

Abstract

Macroeconomic instability, characterised by high inflation, a fragile foreign exchange

position and high rates of interest, increases uncertainty for any investor or producer and
hence slows down economic growth. While this is generally accepted, the usual perception
about the agricultural sector, particularly in India, is that it is immune to general
macroeconomic shocks. In this paper, we intend to examine this perception using a vector
auto regressive model. By studying the significance of macroeconomic conditions to the
agricultural sector, we observe that the sector is not insulated from macroeconomic
shocks.

1. Introduction

A conducive macroeconomic environment is necessary for rapid economic growth.

Macroeconomic instability, characterised by high rates of inflation, a fragile foreign
exchange position, high rates of interest, increases uncertainty for any investor and
producer and hence slows down economic growth. Besides these direct indicators,
macroeconomic instability may also be indicated by overall imbalances such as the fiscal
balance and external current account balance, especially when prices are under
administrative controls. An underlying assumption in these arguments is that production
sectors are influenced by macroeconomic conditions. Any attempt to examine this
proposition will require identification of the indicators of macroeconomic conditions and
of the performance of the sectors.

In the Indian context, the agricultural sector has been important from a policy perspective
for several reasons. Even from the point of view of accelerating economic growth,
transition from an agrarian economy to an industrial or modern economy would depend on
how well the agricultural sector enables this transition. Therefore, besides the concerns
relating to employment and poverty alleviation, the performance of agriculture is of policy
interest from the viewpoint of accelerating economic growth as well. In this context, the
general belief is that overall macroeconomic policies have little effect on the agricultural
sector and that we need sector-specific policies to boost this primary sector of the
economy. While the latter may be true, the macroeconomic environment may also have
non-trivial effects on the agricultural sector. It is necessary to examine this hypothesis
more rigorously as if it does influence agricultural performance, it would be important to
understand the significance of this impact in designing policies for agriculture.

*
Shashanka Bhide is Senior Research Counsellor at National Council of Applied Economic Research, New Delhi; Meenakshi Rajeev is
Associate Professor and B.P. Vani is Assistant Professor at the Institute for Social and Economic Change, Bangalore. The authors wish
to acknowledge the research assistance of Ms. G. Aparna in the preparation of this paper.

1
The key to understanding the impact of changes in macroeconomic parameters at the
sectoral level is the transmission mechanism of these policy impulses to the various
sectors both directly and indirectly through inter-sectoral relationships. Such impact
assessment is often made in the framework of macroeconomic models in which agriculture
and other sectors are featured in detail. Rangarajan (1982) provides one of the early
attempts at estimating the inter-linkages between agriculture and industry. There are other
studies, such as those of Narayana, Parikh and Srinivasan (1991), Strom (1993) and
Kalirajan and Bhide (2003), where the impact of macroeconomic policies on agriculture is
simulated using economy-wide models for India. Shand and Kalirajan (1999) provide
another approach at capturing these linkages where they examine inter-sectoral
dependence through Granger causality tests. Both these approaches capture two-way links:
from agriculture to the non-agricultural sectors and vice-versa within an implicit or
explicit specification of a macroeconomic environment. In one of its recent reports, the
Reserve Bank of India (2002) draws attention to the impact of rising food subsidies on
macroeconomic conditions. However, the analysis is limited to the one-way linkage.

India saw very wide-ranging changes in macroeconomic policies in the 1990s. The
changes at the macroeconomic level were changes in the fiscal and financial sectors, trade
and investment policies. It has been argued that agriculture was only indirectly affected by
these reforms. The industrial sector was most affected directly by the removal of the
production licensing system enforced through control over new investments. Changes in
the financial sector including exchange rate policies and attempts to stabilize fiscal
imbalance, however, can be expected to have an impact across all the sectors. How
important was this impact to agriculture?

The economy-wide models, of macroeconomic variety or the CGE type, provide a

comprehensive analytical structure for analysis. One limitation of the macroeconomic
models is of course that much effort is needed to build the structural equations that
incorporate the various inter-relationships and it is often difficult to check for the impact
of changes in polices due to changes in the structure. A more flexible approach to the
assessment of the impact of impulses emanating from the macroeconomic factors to
agriculture is the framework of time-series analysis. We should note at the outset that the
VAR approach essentially captures the ‘reduced form’ relationships among the selected
variables. Interpreting the estimated linkages in a theoretical framework is not easy
because the theoretical specification is not complete in a VAR.

This paper is an attempt to assess the nature of the inter-relationship between selected
macroeconomic factors and agriculture using the vector auto regression (VAR) approach.
The VAR approach provides a general framework for assessing the impact of inter-related
variablesi. The general framework of analysis we adopt here is to first specify a set of
variables that capture the performance of agriculture and a set of variables that specify the
macroeconomic environment. The agricultural variables we consider are real agricultural
GDP, agricultural exports and fixed investment in agriculture. The last mentioned variable
is measured through the gross fixed capital formation (both public and private) in the
agriculture sector. The macroeconomic variables are interest rate, foreign exchange rate of
the rupee and fiscal deficit of the Central Government. We then use two methods of
quantifying the impact of macroeconomic factors on agriculture. One approach is to
estimate the VAR and the impulse response functions. The second approach is the
variance decomposition to quantify the impact of the macroeconomic factors on
agriculture. We have used annual data for the period 1970-71 to 2003-04 for the analysis.

2
This period covers a variety of experiences both in macroeconomic conditions,
agricultural performance and policies.

The rest of this paper is devoted to the presentation of the results of analysis and
discussion of the findings. To provide a context for the analysis that follows, we briefly
review the trends in some of the major macroeconomic variables and discuss the likely
impact of these changes on agriculture. We then present a discussion of the trends in
selected macroeconomic variables, followed by methodology and results of the VAR
analysis respectively. A concluding section follows at the end.

2. Conceptual Framework

India’s macroeconomic policies have generally attempted to ensure adequate resources for
the investment programmes of the public sector while maintaining adequate supplies of
essential commodities for mass consumption. Sectoral policies, whether in agriculture or
industry, were cast within the overall framework of macroeconomic goals. India’s
macroeconomic stabilisation programme of the 1990s that preceded and then overlapped
the structural adjustment reforms aimed at reducing the fiscal imbalance, moderating
inflation, correcting the overvalued foreign exchange rate and bringing down interest rates.
The emphasis on public sector investment has changed to investment that is commercially
viable. While food security and economic growth remain critical objectives, there is
greater stress on maintaining the conducive macroeconomic environment rather than on
direct public investment at the micro level. Do these changes have an impact on
agriculture? To attempt an assessment of this question, we will need to identify the factors
that describe the macroeconomic conditions and the variables that describe performance of
agriculture. In this section, we identify these factors and variables, examine the trends in
these variables over time and provide a discussion of the potential mechanisms by which
the impact of changes in the macroeconomic conditions is transmitted to the agricultural
sector.

Selection of Variables for Analysis: Of the many indicators of macro-economic

conditions, we focus on aggregate market imbalances and aggregate prices to discern the
impact of macro-economic conditions on agriculture. The imbalance that we have chosen
to reflect macro-economic conditions in this study is the fiscal imbalance1. The three
aggregate prices chosen for the analysis are interest rate, exchange rate and terms of trade.
Prices would reflect market conditions fully only if policy measures are not used to control
these prices in the divergence of market forces. In the Indian context, although controls
over prices existed, at the aggregate level, the controlled prices also saw gradual
adjustments with respect to market conditions. Besides the potential for direct impact,
these variables capturing the macro level imbalances and price conditions also trigger
policies that may have a direct impact on specific sectors. For example, high rates of
inflation bring (captured by terms of trade) to focus the need for ensuring adequate
supplies of essential goods of consumption and therefore greater attention to policies that
raise agricultural growth.

In this sense, the choice of variables to reflect macro-economic conditions should include
the indicators that not only have a direct impact on the performance of agriculture but may
1
We have also considered imbalances that arose in the external current account and did a similar study.
Qualitatively results remain the same (see Bhide, Vani, Rajeev,2005) .

3
also influence agriculture indirectly through policies resulting from macro-economic
conditions.

With these considerations, four broad measures of macro-economic conditions selected for
further analysis in this study are: gross fiscal deficit of the central government (GFDC),
and interest (PLRX), real effective exchange rate (REER) and terms of trade (TT).

To assess the impact of macroeconomic factors or conditions on agriculture, we also need

to identify the variables that reflect the performance of agriculture. In this study, we
consider the variables that capture different dimensions of agricultural sector. Agricultural
investment, agricultural exports and agricultural GDP are the three variables selected in
the study to reflect the performance of agricultural sector. They reflect the overall output
performance of agriculture and also relate more directly to interest rate, inflation rate and
exchange rate changes. Interest rate conditions affect agricultural investment and changes
in exchange rate influence exports. The overall agricultural GDP is inter-linked with
prices, measured through terms of trade, and all other factors that influence either the
demand or the supply of farm products.

In most cases, of the above macro-economic indicators, alternative measures are

available. In the case of interest rate, a number of interest rates are available indicating the
wide range of financial markets. We have selected a rate that is a benchmark for
investment lending by the commercial banks. A combination of the minimum lending rate
of IDBI and the Prime Lending Rate of commercial banks (PLRX) was chosen as the
interest rate for analysis in this study. In the case of exchange rate, we have selected real
effective exchange rate of the rupee (REER), which is a trade weighted real exchange rage
of 36 major trading partners of India. We have used the wholesale price index (WPI) for
agricultural commodities and manufactured products to define a standard measure of terms
of trade (TT) as (WPI for agriculture/WPI for manufacturing)*100.

(a) The Macroeconomic Trends:

The key variables that are tracked in this section (Table A.2 in Appendix presents the data
used for the study) include fiscal deficit and the price related variables. We have focused
here on the fiscal deficit of the Central Government, as the initial correction under the
stabilisation program was at this level of government. In addition, we also present the
trends in the terms of trade, real exchange rate (REER) and interest rateii.

The intense pressures of fiscal and external sector imbalances at the time of the
macroeconomic crisis of 1990-1991 are well known. The macroeconomic crisis triggered
many economic policy changes. Two of the key indicators that reflected the crisis were the
fiscal and external imbalances.

The macroeconomic adjustment process focused attention on the gross fiscal deficit of the
Central Government although the overall fiscal imbalance is known to be much higher
than this deficit. There were similarities in the two measures in terms of direction of
change although the relative changes have been different. In the Appendix we present the
data only for the Central government deficit.

4
Turning to the three price variables in the aggregate markets, the nominal interest rate,
PLRX, increased sharply in 1991-92 and remained about 15% up to 1996-97. Since 1996-
97 there has been a decline in PLRX. In nominal terms, the decline is by almost 4
percentage points since the high levels of 1995-96. The decline is less marked in terms of
real interest rate. However, it must be pointed out that there has been some lending below
the PLR by the commercial banks indicating that trends in PLR can only be a crude proxy
for the trends in interest rates in the economy. The drop in interest rates has been a feature
of the economy since the mid-1990s.

The real effective exchange rate, REER, saw a major correction in 1991-92 and 1992-93
after a steady depreciation for about a decade. Since then there has been a relatively stable
period marked by a tendency towards appreciation. Although controls on external capital
account transactions remain, the rupee is sensitive to supply-demand pressures in foreign
exchange markets. The large levels of forex reserves moving closer to $90 billion have
led to the strengthening of the rupee.

Indicators of terms of trade (calculated as a ratio of wholesale price index for agricultural
commodities to wholesale price index of manufacturing products) reflect a relatively
stable behaviour during our period of study. The index rose sharply between 1975-76 and
1982-83 after which there has again been stagnation up to 1995-’96 thereafter we observe
some decline. The period of stable terms of trade from the mid-1980s to mid-1990s
includes the period when fiscal imbalances were growing and were high, as well as period
of high rates of inflation. We also note that this has been a period when private sector
capital formation in agriculture was rising. The years since 1995-96 up to 2002-03
include a period when non-agricultural investment was also on the decline and overall
inflation rate decreased especially after 1998-99. In other words, agricultural prices have
kept pace with the overall inflation rate especially when the inflation rate has been
relatively high. Macro-economic instability, which included high rates of inflation, was
also characterised by higher growth in agricultural prices.

(b) Performance of Agriculture:

The trends in key agricultural variables in this analysis are illustrated in Figures 1, 2 and
Table A.2. Figure 1 shows the gross fixed capital formation (GFCF) in agriculture both
public and private sector combined. A disaggregated analysis of this data reveals that
private agricultural investment (gross fixed capital formation in constant prices), indeed
rose sharply between 1987-88 and 1990-91, became stagnant for the next three years till
1993-94. It rose steadily again till 1998-99 after which it remained at the same level in the
subsequent year. The years 1991-92 to 1993-94 formed a period when the macro-
economic parameters were unstable reflecting the adjustments in policy. The subsequent
period was marked by a few years of strong growth in industry and a climate favourable
for investment. This period also appears to have influenced private sector investment in
agriculture. However, the public sector capital formation in agriculture has continued to
stagnate for well over a decade. Fiscal pressures on the one hand and preference for
subsidies have led to stagnation in Government spending on investment in agriculture.
The impact of adverse macro-economic conditions on investment may have also affected
in particular public investment in agriculture.

5
Figure 1. Investment in Agriculture (GFCF) (Rs. Crore 1993-94 prices)

25000.0

20000.0

15000.0

10000.0

5000.0

0.0

Year

Note: Rs 1 Crore= Rs 10 million.

Agricultural exports increased sharply during the period 1988-89 to 1996-97 (Figure 2).
Exports in value terms declined to some extent from 1996-97 onwards but again increased
in the current decade. The latter decline is attributed to a decline in the unit value of
exports during the period when global commodity prices also experienced a decline.iii
Macro-economic factors primarily relating to the real exchange rate would have an impact
on exports, including agricultural exports. The decline in exports has occurred during a
period when the real exchange rate has been relatively stable or slightly appreciating.

Figure 2. Agricultural Exports (in Rs crores, 1993-94 prices)

6000.0

5000.0

4000.0

3000.0

2000.0

1000.0

0.0

year

6
T rends in agricultural GDP trend at constant 1993-94 prices show a steady increase over
the years, except for a slight fall in a few years (Table A.3 in Appendix).

3. Methodology for Assessing the Impact of the Macroeconomic Factors

The broad trends in the macroeconomic factors and measures of agricultural performance
indicate fluctuations and changes in pattern over a long period of over three decades
(1970-71 to 2004-05). In the case of macroeconomic variables, the trends reveal upward
and downward movements in fiscal deficit, downward movement of terms of trade,
correction in exchange rate and drop in nominal interest rate. How would these changes
have an influence on agriculture?

The policy channel: The mechanisms through which changes in the macroeconomic
variables are transmitted to agriculture are several. At a general level, macroeconomic
imbalances reflected in the levels of fiscal are characterised by their composition as well.
Lower fiscal deficit may be achieved by expenditure compression or revenue expansion.
The manner in which imbalances are realised may also have an impact of its own. Besides
these composition effects, the ‘twin deficits’ have an impact on aggregate prices: terms of
trade, interest and exchange rates. More importantly, the imbalances also lead to policy
responses such as controls on credit availability, access to markets and quality of
government services each of which affects all producers, including agriculture.

The investment channel: The market signals induced by macroeconomic changes have an
impact on investment. Changes in nominal interest rate and output prices affect real
interest rate, which in turn influences investment decisions of farmers. Further, Changes
in terms of trade may imply differential changes in sectoral price indices, which in turn
effect investment decision of a farmer. In other words, there may be changes in the ‘terms
of trade’ not only due to the changes in prices resulting from structural factors such as the
changes in tariff rates but also due to differences in the speed with which different prices
adjust to the macroeconomic shocks. In addition, poor fiscal conditions also affect
government spending on investment projects.

The export channel: The third mechanism is the impact of changes in exchange rate on
agricultural exports. Changes in real exchange rate influence the competitiveness of
exports in general and hence agricultural exports as well. Changes in the performance of
exports influences total demand for farm output and hence overall agricultural output.

There are, thus, potentially individual or direct effects of changes in macroeconomic

conditions on agriculture and these effects are further influenced by inter-relationships
within agriculture. However, we should also point out that the macroeconomic variables
themselves are inter-related and the impact of one change influences the others and the
impact on agriculture is not limited to just the change in one factor. The Vector Auto
Regressive method is a suitable econometric tool to analyse such inter-relationships.

7
Methodology

The vector auto regression (VAR) representation ( or a Vector Error Correction (VEC)
Model , if the variables are cointegrated) of variables allows an assessment of the inter-
relationship in a dynamic framework. Although this representation has often been termed
‘a-theoretic’, choice of the variables in the VAR can be guided by theory.

The VAR (or, VEC) framework permits us to examine the ‘impulse response functions’
of one of the variables in the VAR to shocks in the other variables. In other words, we are
able to assess the response of say agricultural investment to macroeconomic shocks in a
VAR framework. However, in one of the approaches, the results of VAR analysis would
be sensitive to the ordering of variables in the VAR. To overcome the ambiguity, the
‘generalised impulse response functions’ have been developed (Pesaran and Shin, 1998).

The ‘variance decomposition’ of the VAR allow us to quantify the contribution of

different variables in a VAR to the variability of a selected variable. For example, the
contribution of macroeconomic variables to the variance of agricultural investment
provides an assessment of the impact of the macroeconomic factors on agriculture.

Though the impulse responses measured through ‘generalized impulse response’ approach
are independent of the ordering of the variables, to arrive at variance decomposition one
needs to specify an ordering. Thus it is not possible to completely get rid of the ordering
issue. Further an ordering based on theoretical justification is expected to provide better
indication of the effect of different factors. Keeping this in mind we have considered two
alternatives of sequencing of the variables2. The macroeconomic variables considered in
the present analysis are:

i. (Gross fiscal deficit/ GDP at market prices)

ii. Terms of trade: (WPI agriculture/WPI manufacturing)*100
iii. Interest rate (PLRX: Minimum lending rate of IDBI and PLR of commercial banks)
iv. Real exchange rate (REER based on 36 country trade weighted bilateral rates)

The agricultural variables are:

i. Agricultural investment (Gross fixed capital formation in constant prices)
ii. Agricultural exports (in constant prices)
iii. Agricultural GDP (in constant prices)

The VAR model under consideration is

VAR based on (GFDC, REER, PLRX, GDP_AG, EXP_AG, TT, GFCF_AG)

Proceeding with the analysis,. first for the data under consideration unit root test is carried
out to identity the order of integration using the following test. Three different tests are
applied to attain robustness, viz.,
 Augmented Dickay- fuller test

2
One in the main text and other in Appendix.

8
  Schmidt – Phillips test: two alternatives i.e Z(Rho) and Z ( Tau)
 KPSS test ( kwiatkowski, Phillips, Schmidt & shin(1992) which tests the null
hypothesis of stationarity, level stationarity and trend stationarity.

Results are presented in table1 and they indicate that most of the variables are I(1) series3.

Table 1 Unit Root test result

Schmidt Philips test KPSS

ADF Rho test Tau test Level Trend

stationarity Stationarity
AG_GDP I(1) I(1) I(0) I(1) I(1)
AG_EXP I(1) I(1) I(1) I(1) I(1)
AG_INV I(1) I(1) I(1) I(1) I(1)
REER I(1) I(1) I(1) I(1) I(1)
PLR I(1) I(0) I(1) I(1) I(1)
TT I(1) I(1) I(1) I(1) I(1)

The next step is to check whether or not the variables are cointegrated. If the they are, then
formulating a pure VAR model will be inappropriate. Johansen trace and Saikkonen and
litkepohl tests are used to examine this. Here it should be noted that the number of
cointegration structure, depends on the number of lags in the model. To identify the lag
structure AIC, HQ and Schwatz criteria were used. These tests indicate 3 as the optimal
lag length implying VEC model will have 2 lag structure (results in table 2).

Table 2: Cointegration test result

Johansen trace test Saikkonen and Lutkepohl test

R0 LR p-value R0 LR p-value
0 202.14 0.0000 0 152.52 0.0001
1 109.90 0.1396 1 81.07 0.2137
2 80.53 0.1698 2 53.97 0.3490
3 55.72 0.2009 3 30.94 0.5925
4 34.70 0.2645 4 13.63 0.8507
5 18.72 0.3040 5 10.67 0.2805
6 5.39 0.5506 6 1.30 0.7124

Both the tests suggest the presence of one cointegration vector. Thus VEC model was
constructed with one cointegrating vector and with 2 lags.

The general form of the VEC model under consideration is as follows.

3
Except for a few variables like GDP_AG and PLR follows I(0) under Schmidt Philips test.

9
  Yt −1 
 
Γ 0 ∆ Yt = α [β ′ n′ ]  ....  + Γ 1 ∆ Yt-1 + ……..+Γ p ∆ Yt-p +c Dt + Ut
co
 D t −1 

where, Yt = (Y1t…..Ykt) ′ is a vector of K endogenous variables; Dt contains all

co

deterministic terms ( like Constant, trend etc) included in the co integration relations and
Dt corresponds to the deterministic terms of the VAR component. The parameter α and ß
have dimensions (k× r), where r is the cointegration rank ;β represents the cointegration
parameter and α the loading co-efficients; Γ ’s are the co efficient of the VAR part.

With this specification a VEC model was run and the estimated cointegration relation
turned out follows.

10
1.0 GFDC t-1− 0.146 REERt-1 – 0.022 PLR – 0.690 TT + 0.0001 GDP_AG + 0.002
EXP_AG + 0.004 GFCF_AG + 120.30 = 0

Thus we observe that GDP_AG, EXP_AG and GFCF_AG show positive responses to
change in REER, PLR and TT and negative response to GFDC in the long run. The
agriculture related variables are also positively associated with each other. Thus increase
in agriculture investment will have positive impact on agriculture output. Similarly,
increase in agriculture export will also have positive impact on agriculture output. Further
improvement in terms of trade for agriculture impacts all the agriculture related variables
positively.

The co integrated vector was further checked for stationarity. All the three tests indicated
that co-integrated vector is I(0) series. The next step is to make sure that the errors are free
from auto correlation and the model is stable. The LM test indicated that the residual are
free from auto-correlation. The roots of the reverse characteristic polynomial was found to
be

(1.1242 1.1883 1.1883 1.1693 1.3100 1.9672 2.4676 1.4327 1.4327 1.3140 2.7640
2.7640 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 )

For the condition of the stable VEC model with Ψ endogenous variables and 1 co
integrating equation the companion matrix should have exactly (Ψ -1) i.e., in this case we
should have 6 unit roots and the remaining ones should be strictly greater than one. From
the above numbers we can see that we have exactly 6 ones the rests are all greater than one
indicating that the model being estimated is a stable one. Policy responses further
generated from the model are as follows.

Impulse Response Functions:

VAR methodology provides an estimate of the impact of a shock in terms of change in one
component of VAR on all the components over time. The impact is measured by the
impulse response coefficients (Greene, 1997). We discuss below the results of the impulse
response analysis4. In each of the cases described below, the impact on agriculture related
variables is a result of a one-standard error increase in a selected macroeconomic variable.

Impulse response analysis provides an estimate of the impact of a shock over time
beginning with the year (period) in which the shock is administered. For the purposes of
the present analysis, we have presented the results over a 25-year period from the year of
the shock, however, one can see that substantial impact is only for five years or so. We
note that although the initial shock to the system is in terms of one macroeconomic
variable, in the years following the initial shock, the other macroeconomic variables
respond to the initial shock and in turn influence the agricultural variable. Thus, the
subsequent impact on an agricultural variable, after the initial year, comprises the impact
of the initial shock, secondary impacts from the other macroeconomic variables and the
dynamics of the agricultural variable itself. In the discussion, we attribute the impact to the
4
The estimated VAR equations are not reported in the paper to conserve space. They are available with the
authors.

11
initial shock but it should be borne in mind that secondary influences are at work, besides
the primary or initial shock. Given this analytical background we begin with the first
model specified above:

Agricultural Investment and the Macro-economic Factors: Figure A-1 in Appendix

presents the estimated impact of the one-standard error shock of each macroeconomic
factor (one at a time) on agricultural investment. Some general observations can be made
from the impulse response patterns. First, The impact of shocks given to most of the
variables is maximum during the first 5 years but some impacts can be seen in the later
years as well. However, after 10 years or so the effects become minimal and the responses
stabilizes. Second, fiscal deficit appear to have negative impact on agricultural investment
in the long run even though some of these deficits may have been incurred in order to
subsidise agriculture. As expected an increase in interest rate measured through PLRX has
negative impact ; while increase in terms of trade has a positive cumulative impact. The
impulse responses are also summarised in Table 2 below for the impact in the year 1 of the
shock (Year 1 in the table), in two subsequent years and cumulative impact for 5, 10 and
25 years.

Table 2. Results from Impulse Response Analysis: Impact of Macroeconomic Shocks

on Agricultural Investment*
Cumulative impact
st nd rd
1 year 2 year 3 year 5 yrs 10 yrs 25 yrs
GFDC -34.90 16.42 -124.14 -80.67 -87.43 -83.50
REER -61.01 -114.42 -285.93 -280.24 -198.44 -231.58
PLR -0.29 -139.35 46.08 -150.85 -160.14 -142.46
TT 251.35 -80.90 122.11 170.09 147.00 100.71

*measured in crores of rupees (10 million rupees) at constant prices.

A rise in the real value of the rupee (rise in REER) leads to a decline in agricultural
investment in terms of machineries.

Agricultural Exports and Macro-economic Factors: The impulse responses of

agricultural exports are presented in Figure A.1-2 in the Appendix. Although fiscal deficit
has fluctuating impact in the initial years (see the first three years in Fig. A.1-2) the overall
long run impact is negative. The initial (first period) impact on agricultural exports is
positive for an increase in the interest rate (PLRX). In the later period, there is a larger
negative impact, although the subsequent impacts are not uniform over time but much less
fluctuating. The impulse response is summarised in Table 3 below for agricultural exports
and also in Figure A.1-2 in Appendix. An increase in agricultural price (TT) has positive
impact on agricultural export. We have seen earlier that TT also has positive impact on
agricultural investment. Thus favourable terms of trade are having positive impacts on the
sector. Given the impact of price increase on agricultural export, an increase in REER may
capture an appreciation of the value of rupee. Usually this should result in negative impact
on export. However, an appreciating rupee may in turn imply an overall good
macroeconomic condition and hence a positive impact on export, which we observe in this
situation.

12
Table 3. Results from Impulse Response Analysis: Impact of Macroeconomic Shocks
on Agricultural Exports*

Cumulative Impact
1st year 2nd year 3rd year 5 yrs 10 yrs 25 yrs
GFDC -111.51 -44.28 2.66 -135.59 -127.75 -133.48
REER 51.38 42.08 14.52 124.96 121.15 119.75
PLR 5.58 52.96 -118.88 -111.95 -89.64 -80.60
TT 102.64 8.18 30.05 102.95 89.57 100.67

* Measured in U S million dollars

Agricultural GDP and Macroeconomic Factors: The overall performance measure for
agriculture selected in this analysis is agricultural GDP. Impact of macro-economic
shocks on agricultural GDP captured in the impulse responses is spread over relatively
short periods of time (Figure A.1-3 in the Appendix). We observe that an increase in
fiscal deficit has a negative impact on agricultural GDP. Increase in fiscal deficit is earlier
seen to effect both investment and export also adversely. This may be due to adverse terms
of trade resulting from inflationary pressure created by fiscal deficit. The impacts of
REER and PLRX also have expected sign. The results are summarized in Table 4.

Table 4. Results from Impulse Response Analysis: Impact of Macroeconomic

Shocks on Agricultural GDP*

IRF: Agricultural GDP

Cumulative Impact
rd
1st year 2nd year 3 year 5 yrs 10 yrs 25 yrs
Fis_def -2453.28 -1046.98 2407.88 -2487.44 -2450.19 -2140.02
REER -2400.45 3430.56 -2337.67 -1581.32 -1077.04 -1219.86
PLR -1403.88 566.44 -443.39 -1090.63 -1476.31 -1346.92
TT 1101.25 1016.06 .2239.24 2457.82 1488.18 1112.97

* Measured in crores (10 million) of rupees at constant prices

Integrated View of the Results: Results from three different specifications of the VARs
show that the impact of different macroeconomic factors on the three selected agricultural
performance variables are quite similar only except in the case of effect of REER on
agricultural export. For example, an appreciation of REER is seen to affect agricultural
exports and agricultural investment differently. First is to note than exports constitute only
a small part of agricultural output. Secondly export decisions are based on entirely
separate set of considerations. From variance decomposition also it can be seen that REER
explains only 2.8% of variability of agricultural export (see Table 6). To provide an
integrated view of the results, we summarise in Table 5 the estimated impact from the
VAR by aggregating impulse responses over 5 year period. It should be noted here that
qualitatively the integrated picture does not change even when we consider impulse
responses over 25 or 10 year period.

13
Table 5. Impact of the Macroeconomic Factors on Agriculture: An Integrated
picture (Cumulative Impulse Response over 5 years)
Macroeconomic Agricultural GDP Agricultural Agricultural
variable exports investment
GFDC Negative Negative Negative
PLRX Negative Negative Negative
REER Negative Positive Negative
TT Positive Positive Positive
Note: The results that are different from the remaining two cells across the row are in bold
font.

Table 5 shows that in almost all cases we observe expected linkages, except in the case of
REER, which shows a positive impact on export.

Clearly, it is difficult to identify structural relationships through VARs. But the results
presented above show that the macroeconomic factors do influence agriculture through a
number of interfaces. Thus, the agricultural sector is not immune to macroeconomic policy
changes as generally perceived.

(c) The Variance Decomposition:

VAR methodology provides a means of assessing the contribution of different variables in

a system to changes in any given variable within the VAR. The ‘variance decomposition’
component provides an estimate of the contribution of a variable to the variability of
another variable in the VAR. We apply variance decomposition technique to assess the
contribution of the macroeconomic variables to variability in the selected agricultural
variables. Table 6 summarises the results of the variance decomposition analysis.

The results indicate that many of the macroeconomic factors are important in explaining
the variability in agricultural performance indicators. In fact The VAR results show that
40--50% of variability in agricultural performance indicators is captured by the macro-
level factors. This is a strikingly significant impact and indicates that agricultural sector is
far from being insulated from the overall macroeconomic conditions of the economy. In
particular, the macroeconomic factors in different VARs also point to the fact that the
fiscal deficit has significant contribution in explaining the variability in the case of all
three agricultural variables. Amongst the price related variables impact of interest rate
appears to be minimal. As expected agricultural GDP has significant impact on
agricultural investment and export. Further past values of the variables explains about
50% (or more) of the variability in all the cases.

Table 6. The Impact of Macroeconomic Factors on Agriculture: Percentage of

Variance in Agricultural Performance Variables Explained by Macro-economic
Factors

GFDC REER PLAR TT -80.90 AG_EXP AG_INV

AG_GDP 4.6 8.1 1.1 5.4 66.9 3.3 10.0
AG_EXP 7.6 2.8 9.0 7.2 11.0 51.3 11.0
AG_INV 1.0 6.2 1.8 4.9 8.8 7.5 69.7

14
5. Policy Implications and Concluding Remarks

In this paper, we set out to assess the impact of macroeconomic factors on agriculture
using the time-series approach. There are clearly several mechanisms by which the shocks
to macroeconomic variables would be transmitted to decision variables in agriculture and
finally affect the performance of the sector. For the purposes of the present analysis, price
variables, viz., exchange rate, terms of trade and interest rate and the macro imbalances,
captured through gross fiscal deficit of the centre to GDP are considered. We also chose
three agriculture-related variables: agricultural investment (real), agricultural exports (real)
and real agricultural GDP for measuring the sensitivity of agricultural sector to the
macroeconomic conditions.

Results of impulse response analysis capture the direction of the impact of macroeconomic
shocks to agriculture. The direction of the impact is generally along the ‘expected lines’
indicated by structural relationships, suggesting significant inter-relationships of variables.
First it indicates that fiscal deficit incurred by providing subsidies to agricultural sector in
the long run does not generate either higher production or export of the sector. This shows
that incurring fiscal deficit to provide subsidy or other stimuli to the sector may not be
very effective in the long run. On the other hand relatively higher price for agricultural
output may induce the farmers to invest, export and produce more. Thus, providing
necessary security to the farmers through remunerative procurement prices and ensuring
better access to the market (without incurring fiscal deficit) are essential for the sector.

Interestingly, improvement in the value of rupee does not have an adverse effect on
export. Improvement in the value of rupee witnessed in recent years is due to the general
opening up of India’s markets resulting in greater inflow of foreign exchange. Liberalized
markets have also helped improve export potential of all sectors including agriculture.
Therefore, even with an appreciating ‘REER’ one can observe improvement in exports.
However, this increase in exports is not accompanied by an increase in agricultural output
or investment when the source of export is appreciating rupee. This could be due to the
fact that appreciating rupee also induces imports, which may meet some of the domestic
consumption needs. Furthermore, there may also be other inter-relationships that are not
considered here explicitly.

Another interesting finding is that interest rate has comparatively much less impact on the
agricultural variables especially investment and output. The presence of excess demand
for formal credit and the demand- rationed nature of the credit market are possible
explanation for this outcome. This is clear from the wide spread existence of informal
sector lenders. If a change in the formal interest rate is accompanied by a corresponding
change in the supply of credit then one may expect to see a significant impact. However,
given the rigidity of supply of credit from the formal credit supply channels, the impact of
a change in the interest rate on the use of credit is also muted. Thus, it is the availability
and delivery mechanism of credit that plays a more important role the than interest rate.

The present analysis has pointed to the substantial impact of macroeconomic factors on
agriculture. The analysis does not track all the transmission mechanisms but points to the
likely links. While the results need to be qualified by the underlying assumptions, it is
difficult to ignore the need to keep in view the macroeconomic factors in understanding
the changes in the agricultural sector.

15
 References

Bhide, S, M Rajeev and B P Vani, 2005, Do Marcoeconomic Conditions Matter for

Agriculture? The Indian Experience, Working Paper Number 162, Institute for Social and
Economic Change, Bangalore, India

Economic Research Foundation, 2002, National Accounts Statistics, CD, Mumbai.

Enders, W. 1995, Applied Econometric Time Series, John Wiley and Sons, Inc., NY.

Greene, W.H. 1997, Econometric Analysis, Third Edition, Prentice Hall, Englewood Cliffs
NJ.

Kalirajan, K.P. and S. Bhide, 2003, A Disequilibrium Macroeconometric Model for the
Indian Economy, Ahgate, London.

Narayana, N. S. S., Parikh K. S. and T. N. Srinivasan, 1991, Agricultural Growth and

Redistribution in Income: Policy Analysis with a General Equilibrium Model of India,
Elsevier Science Publishers, Amsterdam.

Pesaran, H.H. and Y. Shin, 1998, Generalized Impulse Response Analysis in Linear
Multivariate Models, Economics Letters, 58, pp. 17-29.

Rangarajan, C, 1982, Agricultural Growth and Industrial Performance in India,

International Food Policy Research Institute, Washington, D.C.

Reserve Bank of India, 2002, Report on Currency and Finance, 2001-02, Mumbai.

Reserve Bank of India, 2003, Handbook of Statistics on Indian Economy, 2002-03,

Mumbai.

Shand, R. and K.P. Kalirajan, 1999, The Agriculture-Manufacturing Nexus and

Sequencing in India’s Reform Process, in editor, Economic Liberalisation in South Asia,

16
R. Shand (ed.) Macmillan India Limited, Delhi.

Sims, Christopher, 1980, Macroeconomics and Reality, Econometrica, 48, pp.1-49.

Strom, S., 1993. Macroeconomic Considerations in the Choice of an Agricultural policy:

A Study into Sectoral Interdependence with Reference to India, Ashgate Publishing group,
London.

17
 Appendices
Appendix A.1: Details of VAR analysis

Table A.1.1 Unit Root Tests of Selected Variables

Variable ADF Statistic for Variable in Order of
Level form Level form First First Second Second Integration
lag1 lag2 Difference Difference Differenc Difference
form lag1 form lag2 e form form lag2
lag1
REER -1.510 -1.254 -3.886** -3.037 I(1)
(DW 1.936) (DW 1.818) (DW (DW
1.836) 1.451)
PLR -1.112 -0.687 -4.916*** -4.431*** I(1)
(DW 2.008) (DW 2.097) (DW (DW
2.143) 1.938)
GDP_AG -2.618 -1.947 -5.816*** -3.799** I(1)
(DW 2.086) (DW 1.941) (DW (DW
1.980) 1.957)
EXP_AG -1.623 -1.463 -2.1792 -1.508 -4.303 ** -4.026** I(2)
(DW 1.552) (DW 1.546) (DW (DW (DW (DW 1.908)
1.563) 1.635) 1.871)
GFCF_AG -1.797 -2.055 -3.748** -3.061 I(1)
(DW 1.852) (DW 2.030) (DW (DW
1.970) 1.984)
GFDC/GDP -1.372 -0.814 -5.298*** -4.308*** I(1)
(DW 2.028) (DW 1.882) (DW (DW
1.925) 2.018)
TT -3.161 -1.246 -7.567*** -4.466*** I(1)
(DW 1.644) (DW 2.135) (DW (DW
2.108) 1.895)

Note: See text for details of variables

Figure A.1.1 Impulse Response Analysis for Agricultural Investment

Response of agriculture investment to one s.d. real effective exchange rate innovation

18
 Fis_def to AG_INV

100

50

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

-50

-100

-150

Response of agriculture investment to one s.d. real effective exchange rate innovation

REERtoAG_INV

300

200

100

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

-100

-200

-300

-400

REERtoAG_INV

300

200

100

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

-100

-200

Figure A.1.1 Impulse Response Analysis for Agricultural Investment (Continued)

-300

-400

Response of agriculture investment to one s.d. interest rate innovation

19
 PLRtoAG_INV

100

50

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

-50

-100

-150

-200

Response of agriculture investment to one s.d. terms of trade innovation

TTtoAG_INV

300

250

200

150

100

50

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

-50

-100

-150

Figure A.1.2 Impulse Response Analysis for Agricultural Exports

Response of agriculture exports to one s.d. real effective exchange rate innovation

20
 REERto AG_EXP

60

50

40

30

20

10

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

-10

-20

-30

Response of agriculture exports to one s.d. fiscal deficit innovation

FIS_DEF to AG_EXP

40

20

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

-20

-40

-60

-80

-100

-120

Figure A.1.2 Impulse Response Analysis for Agricultural Exports (Continued)

21
Response of agriculture exports to one s.d. interest rate innovation

PLRtoAG_EXP

80

60
FIGURE A-2 (continued)
40

20

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

-20

-40

-60

-80

-100

-120

-140

Response of agriculture exports to one s.d. terms of trade innovation

TT to AG_EXP

120

100

80

60

40

20

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

-20

-40

-60

22
Figure A.1.3 Impulse Response Analysis for Agricultural GDP

Response of agriculture GDP to one s.d. real effective exchange rate innovation

REERto AG_GDP

4000

3000

2000

1000

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

-1000

-2000

-3000

Response of agriculture GDP to one s.d. PLR innovation

FIS_DEFtoAG_GDP

3000

2000

1000

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

-1000

-2000

-3000

23
Figure A.1.3 Impulse Response Analysis for Agricultural GDP (Continued)

Response of agriculture GDP to one s.d. PLR innovation

PLRtoAG_GDP

1000

500

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

-500

-1000

-1500

-2000

Response of agriculture GDP to one s.d. terms of trade innovation

TTtoAG_GDP

3000

2000

1000

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

-1000

-2000

-3000

24
Appendix A.2

Table A .2 Basic Data Used in the study

Agricultural Investment
Exports (Rs Agricultural in Current Fiscal Real
Crores in GDP (Rs Agriculture Account deficit of Effective
1993-94 crores in (Rs Crores Balance the central Prime Exchange
prices) 1993-94 in 1993-94 (% of Government Lending Rate Terms of
Year prices) prices) GDP) (% of GDP) Rate (%) (Index) trade
1970-71 1005.3 7902.0 137320.0 -0.97 3.08 8.5 125.36 97.55
1971-72 1062.9 8349.0 134742.0 -1.02 3.53 8.5 121.78 106.43
1972-73 1212.2 8831.0 127980.0 -0.58 4.04 8.5 119.56 107.79
1973-74 1257.1 8760.0 137197.0 1.73 2.64 9.0 127.50 97.74
1974-75 1456.3 8212.0 135107.0 -1.23 2.97 10.3 114.14 96.96
1975-76 1654.7 8924.0 152522.0 -0.21 3.64 11.0 106.27 106.13
1976-77 1628.7 11066.0 143709.0 1.00 4.24 11.0 101.34 107.79
1977-78 1692.4 11347.0 158132.0 1.11 3.62 11.0 100.12 99.99
1978-79 1582.5 12780.0 161773.0 -0.22 5.18 11.0 91.98 101.84
1979-80 1775.5 13344.0 141107.0 -0.46 5.29 11.0 97.08 111.60
1980-81 1719.6 13721.0 159293.0 -1.54 5.77 14.0 104.48 119.20
1981-82 1753.1 13407.0 167723.0 -1.68 5.14 14.0 104.48 111.60
1982-83 1645.4 13766.0 166577.0 -1.74 5.64 14.0 101.17 107.64
1983-84 1520.5 13926.0 182498.0 -1.51 5.94 14.0 104.24 100.93
1984-85 1578.8 13846.0 185186.0 -1.17 7.09 14.0 100.86 101.49
1985-86 1597.5 13061.0 186570.0 -2.14 7.86 14.0 98.27 107.53
1986-87 1672.6 12789.0 185363.0 -1.87 8.47 14.0 90.24 100.97
1987-88 1643.1 13375.0 182899.0 -1.78 7.63 14.0 85.36 95.53
1988-89 1498.4 14335.0 211184.0 -2.75 7.34 14.0 80.41 98.99
1989-90 1670.7 12728.0 214315.0 -2.34 7.33 14.0 78.44 107.88
1990-91 1768.5 15805.0 223114.0 -3.05 7.85 14.5 75.58 102.87
1991-92 2054.1 14546.0 219660.0 -0.34 5.56 19.0 64.20 95.86
1992-93 2111.5 15610.0 232386.0 -1.71 5.37 18.0 57.08 98.54
1993-94 2526.7 14749.0 241967.0 -0.42 7.01 16.0 61.59 100.00
1994-95 2654.6 15978.0 254090.0 -1.05 5.70 15.0 66.04 96.70
1995-96 4159.5 16824.0 251892.0 -1.65 5.07 16.5 63.62 96.70
1996-97 4515.2 17009.0 276091.0 -1.19 4.88 14.8 63.81 91.20
1997-98 4038.7 17035.0 269383.0 -1.37 5.84 14.0 67.02 91.20
1998-99 4114.4 16516.0 286094.0 -0.95 6.45 12.5 63.44 85.00
1999-00 4010.6 18082.0 286983.0 -1.04 5.35 12.3 63.29 86.20
2000-01 4337.7 18602.0 286666.0 -0.54 5.32 11.5 66.53 86.60
2001-02 4547.1 19402.0 304666.0 0.3 5.07 11.5 68.43 85.10
2002-03 5232.9 19280.0 283393.0 1.24 5.89 10.2 72.76 84.5
2003-04 5010.9 21628.0 310611.0 2.32 4.47 8.9 74.14 85.6
2004-05 5144.2 19428.0 314180.0 -0.8 4.03 8.5 74.53 88.5

25
26
 i
ii

iii