Universal Journal of Agricultural Research 10(4): 405-416, 2022 http://www.hrpub.

org
DOI: 10.13189/ujar.2022.100410

Impact of Climate Change and Economic Activity on

Philippine Agriculture: A Cointegration and
Causality Analysis
Jennifer Madonna G. Dait

College of Business and Management, Ifugao State University, Philippines

Received June 5, 2022; Revised July 28, 2022; Accepted August 20, 2022

Cite This Paper in the Following Citation Styles

(a): [1] Jennifer Madonna G. Dait , " Impact of Climate Change and Economic Activity on Philippine Agriculture: A
Cointegration and Causality Analysis," Universal Journal of Agricultural Research, Vol. 10, No. 4, pp. 405 - 416, 2022.
DOI: 10.13189/ujar.2022.100410.
(b): Jennifer Madonna G. Dait (2022). Impact of Climate Change and Economic Activity on Philippine Agriculture: A
Cointegration and Causality Analysis. Universal Journal of Agricultural Research, 10(4), 405 - 416. DOI:
10.13189/ujar.2022.100410.
Copyright©2022 by authors, all rights reserved. Authors agree that this article remains permanently open access under the
terms of the Creative Commons Attribution License 4.0 International License

Abstract Climate change impact is particularly severe Keywords Climate Change, Agriculture, Time-Series,
in developing countries like the Philippines mainly because Regression
of low incomes, geographic state or condition, dependence
on climate-sensitive sectors and inadequate capability to
adapt to global warming. This paper aimed at analyzing the
risk posed by climate change using climatic variables on
Philippine agriculture. Likewise, it focused on the 1. Introduction
empirical measurement of the hypothesized relationship
between agricultural output and the condition or predicted Climate Change is one of the most alarming incidences
economic variables. This paper employed not only the fronting the world as it also serves as a threat not only to
Cobb-Douglas production function using time series data the environment but also to both social and economic
from 1980 to 2014 but also the modeling techniques - sectors. In the past decades, the increase in global mean
Cointegration and Granger causality to simulate the impact surface temperature has been observed, along with the
of changes of the aforementioned variables on output of increase of average sea level worldwide and melting of ice
Philippine agriculture. Results show that only three in the Northern Hemisphere. Changes in global
variables indicated considerable significance on temperature have contributed much to the volatility of
agricultural production in the Philippines based on their climatic conditions in different areas around the globe [19,
respective t-ratios: Agricultural Employment (EA), 21].
temperature (TEMP), and La Nina (D1). Other things The variability in climate is attributed to various ocean
equally, a 1% rise in agricultural employment paves a 0.2% oscillations that influence the amount of warming or
increase in agricultural production. On the other hand, a cooling – generally known as El Niño (warming) and La
1% increase in temperature, cet. par., decreases agricultural Niña (cooling) (Impact Forecasting 2013). Aside from that,
production by 0.08%. Correspondingly, the incidence of El greenhouse gases (GHGs) from carbon dioxide contribute
Nino, other things equally, deceases agricultural output by to global warming [28].
0.02%. The other variables are not statistically significant Deviations from the current conditions in the mean
but are interpreted in the same way. With this, government climate possibly will entail modifications to present-day
expenditure should be redirected toward R&D in agricultural practices to be able to sustain productivity, and
agriculture to improve resilience, competence and in some cases the optimum type of farming may change
sustainability of the agriculture sector. [14].
406 Impact of Climate Change and Economic Activity on Philippine Agriculture: A Cointegration and Causality Analysis

Theoretical and empirical investigation on predicting and 12% respectively [30]. The agricultural sector’s
climate change effects on agriculture indicates that contribution to the country’s GDP is vital in achieving
different regions vary considerably in whether they win or inclusive growth and poverty reduction.
lose in agricultural suitability, even if average change To supplement changes in the agricultural sector, the
across the entire study region is small. A positive Department of Agriculture has also delivered more
relationship between the wealth of nations and change in infrastructures, farm mechanization, post-harvest facilities,
agriculture conditions was found, but variability around research and extension, and market development [7]. In
this trend was high. Parts of Africa, Europe and southern 2014, it launched the Philippine Rural Development
and eastern Asia were predicted to be particularly Project to lay agri-fishery infrastructure or modernization
negatively affected, while northeastern Europe, can expect to support value-adding activities to enable farmers and
more favorable conditions for agriculture [4]. fishers to participate in growing the economy and share in
The variability in climate affects man’s primary source the gains of development. Aside from that , the Department
of food resource - the environment. Climate change of Agriculture, under the General Appropriations Act of
endangers agricultural productivity and destabilizes 2013, also appropriated a P1 billion loan funds to the
production efficiency. Due to low resilience to climate Agricultural Credit Policy Council (ACPC) to establish a
change and variability, a lot of developing countries across flexible credit facility for the benefit of small farmers
the globe who are agriculture-dependent and insufficiently registered in the Registry System of Basic Sectors in
productive are estimated to have lower agricultural Agriculture (RSBSA) as an alternative to the rigid and
production and capacity [5, 12, 20]. stringent credit facilities usually provided by banks [8].
Increased floods and droughts, soil degradation, water The Philippine government’s program on increasing
shortages, and possible increases in damaging pests and agricultural productivity has increased demands on the
diseases are all possible consequences of climate change. sector’s capacity and has put pressure on its natural
The fishing sector is negatively impacted by hotter water as resource base. The employment of unsustainable
a result of rising temperatures. It has an impact on coral agricultural practices to increase production led to
reefs, causing coral bleaching and poor reef health. Hotter devastation of land resources and water shortage.
water is linked to increased sedimentation and wave action, Unfortunately, climate change has aggravated the essential
both of which harm sea grasses and result in population weaknesses of the sector.
decline. The loss of arable land is caused by rising sea Ample literature and studies are found on the harmful
levels. It promotes seawater intrusion into groundwater effects of climate change on agriculture, however, there are
resources and intensifies soil salinity in low-lying limited studies on empirical climate science using decadal
agricultural lands, causing coastal habitats to drown. variability on climate in the Philippines. Thus, this paper is
Reduced oxygen levels could be a result of climatic change imperative to understand and analyze the linkage between
in water quality, which would lead to fish extinction. This climate and selected economic variables to agricultural
is in addition to ocean acidification, which has an impact production in the Philippines to allow policy-makers in
on mariculture and lowers production [11, 27, 31]. mitigating climate change and develop measures to
Severe changes in climate conditions have brought about increase agricultural production in the country.
environmental damage as a result of floods, droughts, The purpose of this paper therefore, is to quantitatively
denuded forests, soil erosions and loss of endangered evaluate how climate variability and various economic
species; which eventually leads to slow economic growth activities in agriculture affect output of Philippine
because of scarce resources [1]. agriculture. At the same time, it aims to understand the
Though most of the countries have developed mitigation movement of climatic indicators such as temperature,
plans on climate change, IPCC revealed that climate precipitation, number of typhoons that hit landfall, carbon
change has become particularly graver in developing emissions, el niño and la niña in the last thirty-four (34)
countries mainly because of low incomes, geographic state years in so far as the Philippines is concerned; and
or condition, dependence on climate-sensitive sectors and understand the movement of economic factors such as
agricultural credit, agricultural expenditure, employment
inadequate capability to adapt to global warming and the
in agriculture and land area devoted in agriculture in the
Philippines is one of the countries considered highly
last thirty-four (34) years as well. Furthermore, it aims to
vulnerable to climate change [20, 21]. Small-farmers are
find a causal link between agricultural production in the
also the most prone to the damaging effects of the El Niño
Philippines and the aforecited climatic indicators in
Southern Oscillation in the country (ENSO) because of
particular.
limited mitigation policies in the country [16, 33].
Agriculture still plays a vital role in the Philippines’
overall economic development as it provides food and vital
2. Materials and Methods
raw materials for the rest of the economy. Agriculture
employment shares 31% in the total employment ratio, its On understanding the impact of climate change on
total share in the GDP and exports of the country is 10% agricultural productivity, a lot of studies used computable
 Universal Journal of Agricultural Research 10(4): 405-416, 2022 407

general or partial equilibrium models, production functions, (+) (+) (+) (+) (-) (-) (-)
and even Ricardian models, to modify, apply and analyze VAA = f (EA, AE, LA, EA, LN, EN, NT,
the impact of climate variability of precipitation,
temperature and carbon on agricultural production [1]. On (-) (-) (-)
measuring climate change, studies made use of the Prec, Temp, CE) (2)
Seleaninov index [18]. Others made use of the biophysical
statistical model to link the primary climate change impact The algebraic signs on top of each predictor variable
on temperature and precipitation to changes in yield per represent the expected impact of a unit change on one of
unit of land [36]. This method however, is founded on the these conditioning variables, holding other variables
correlation analysis and not necessarily on causal methods. constant on agricultural output. Other things constantly, an
To address this concern, this research used time series increase in capital investments or number of agricultural
models to simulate the long- run and causality relationships workers is likely to cause agricultural output to increase but
between climate change variables, economic activities in the increase would be at a diminishing rate. Increases in
agriculture and value added of agriculture. This paper will number of farmers, agricultural loans, public investment in
make use not only of the Cobb-Douglas function but also agriculture and price index of agriculture will lead to
of the modeling techniques - Cointegration and Granger increased production. On the other hand, increased
causality to simulate the impact of changes in the total quantities of carbon emissions, precipitation and increases
levels of carbon emissions, precipitation, temperature, in temperature based on mounting scientific evidence can
number of typhoons that hit landfall and the incidence of have adverse effects on agricultural production.
the El Niño and La Niña phenomenon on the added value This paper focused on agricultural credit, agricultural
of Philippine agriculture. In the same way, to understand expenditure, employment in agriculture, land area in
the relationship of the agricultural credit, agricultural agriculture, carbon emissions, number of typhoons that hit
expenditure, employment in agriculture and land area is landfall, precipitation, temperature, La Niña and El Niño
devoted to agriculture on the output of Philippine affect total output of Philippine agriculture using
agriculture. regression analysis, thereby creating mitigation measures
The Cobb-Douglas production function or aggregate in combatting climate change.
production function allows determining the output of an
economy given inputs of capital, labor, human capital, and Data Sources and Description
technology. It assumes that the household’s agricultural
productivity (Y) in any time period is a function of In an attempt to determine the relationship between
agricultural Labor input (L), materials (M), physical capital variables of climate change, economic activity on
investment (K), human capital (H) and physical resource agriculture and the output of Philippine agriculture, this
endowment (R). More generally, this may be expressed in study employed annual time series data for the Philippines
the abstract form as follows: from 1980 to 2014. Value added of agriculture was
Y = F (K, L, M, .): F1, F2, F3 > 0; and F11, F22, F33 < 0 (1) gathered from the Bureau of Agricultural Statistics. On the
other hand, data on temperature, precipitation, number of
Where - Y is output, K is usage of capital, L is typhoon that had hit landfall, the El Niño & La Niña
employment of labor, M is the use of raw materials. Other phenomena and carbon emissions were gathered from the
input factors are admittedly possible as described by the Climatic Research Unit at the University of East Anglia,
ellipsis [26]. A similar representation of a generalized the Philippine Atmospheric Geophysical and Astronomical
production function was expressed in a book on managerial Services Administration (PAG-ASA) and the World Bank.
economics [10]. The notations F1, F2, F3 are first order While data on agricultural credit, agricultural expenditure,
derivatives that represent marginal productivities of factor employment in agriculture and land area devoted to
inputs. While the 2nd order partial derivatives F11, F22, F33 agriculture were from the Department of Agriculture and
recognize diminishing marginal productivities for each the Bureau of Agricultural Statistics.
factor.
This paper adopted the standard material-augmented
Empirical Model
Cobb-douglas production function model using the
different variables under investigation: Value added of The economic model describing how agricultural output
Agriculture (VAA) in any time period is a function of is affected by the identified predictor variables followed a
temperature (Temp), precipitation (Prec), carbon emission Cobb-Douglas form. The estimating equation is in
(CE), Number of Typhoons that hit landfall (NT), El Niño logarithmic form as follows:
(EN) and La Niña (LN), employment in agriculture (EA),
agricultural expenditure (AE), and land area devoted to lnVA = ß1 + ß2lnAC + ß3lnAE + ß4lnNF + ß5lnLA +
agriculture (LA). In functional terms, the model will be ß6lnT + ß7lnP + ß8lnCE + ε (3)
described as follows:
408 Impact of Climate Change and Economic Activity on Philippine Agriculture: A Cointegration and Causality Analysis

The Cobb-Douglas production function was used in this cointegration between X and Y. Granger causality analysis
paper to represent the technological relationship of two must be used to demonstrate causality between the
inputs, labor and capital, and the expanse of produce using variables because cointegration between variables does not
those inputs. This function was developed and tested from necessarily suggest that there is a causal relationship
1927 to 1947 by Charles Cobb and Paul Douglas [15]. between them. If X Ganger-causes Y, then X's previous
values can be used to estimate Y's present value [9], hence:

Cointegration Test Yt = ∑ αiYt-1 + ∑ ßjXj-1 (5)

The lagged values of X should not be included in the
Cointegration analysis was used in this paper to
equation if X does not Granger-cause Y.
determine the long run, dynamic relationship between the
predictor variables used in this study. Albeit two variables Ho: X does not Granger-cause Y
are non-stationary individually, it is nevertheless possible a or Ho: ß1 = ß2 = … ßn = 0
stationary linear combination of the two variables [27]. If
Granger Causality testing will be done because, despite
so, X and Y are cointegrated and have a tendency to move
together [10]. This implies that a cointegration test must be the fact that two variables are known to be related, it is
conducted before doing a causality test. To conjure the unclear which variable changes the other. As a result, the
dynamic relationship between the aggregate value of equation: will also undergo a Granger Causality Test in this
agriculture, variables of climate change, and agricultural study.
economic activity, Johansen's Variance Autoregressive Xt = ∑ Yi Xt-1 + ∑ δj Yt-j
(VAR) technique was utilized. [22].
Let the added value of Philippine agriculture be Qt; and
Wt be the seasonal level of the climate variable, t denoting 3. Results and Discussions
time. If there is an equilibrium, long-term relationship
between Qt and Wi, the two time series variables with The outcome of the regression results in the succeeding
stochastic trends are said to be cointegrated [32]. The pages of this paper shows that the model is consistent with
following form can be used to assess whether there is a expectations made by several researches and studies; and
cointegration relationship between Qt and Wi using therefore, was used in the analysis of the threats posed by
Johansen's method: climate change on Philippine agriculture.
y = µ + A1 – Yt-1 + A1 – Yt-1 + . . . + An – Yt-n + εi (4)
Unit Root and Cointegration Test Results
Assuming a VAR process, Yt = (Qt and t Wi) is a n x 1
vector of variables, where µ is a vector of intercept terms, Results show that all the variables were nonstationary at
Yt-i are the lagged values of Yt, Ai are vector coefficient levels as shown in Table 1. All of the economic and
matrices, and εi is a n x 1 vector of error terms. climatic variables showed stochastic patterns, thus the
The standard Granger causality test will be used if the ADF's constant/intercept option was used to assess all of
results of the stationary tests indicate that the two variables, the univariate series at first difference. At the 1 percent
X (variables of climate change and economic activities in level of significance of the MacKinnon critical values, it
agriculture) and Y (aggregate value of Philippine was determined that both the dependent and independent
agriculture) are stationary. variables were stationary. Hence, the variables of the
Granger causality is implied by the existence of model can be regressed at level series.

Table 1. Augmented Dickey Fuller (First Difference, Intercept)

UNIT ROOT TEST AT FIRST DIFFERENCE, INTERCEPT

VARIABLES ADF TEST STATISTICS CONCLUSION
Agricultural Expenditure -7.948037 Stationary
Agricultural Credit -5.373256 Stationary
Employment in Agriculture -9.218647 Stationary
Land Area Devoted in Agriculture -4.282420 Stationary
Carbon Emissions -.3774851 Stationary
Temperature -7.082855 Stationary
Precipitation -5.770639 Stationary
Number of Typhoons that hit Landfall -6.249265 Stationary
Value Added to Agricultural Production -5.899651 Stationary
 Universal Journal of Agricultural Research 10(4): 405-416, 2022 409

Table 2. Cointegration Results

Unrestricted Cointegration Rank Test (Trace)

0.05
Hypothesized Number of CE(s) Eigenvalue Trace Statistics Prob.**
Critical Value

None* 0.858577 178.0447 125.6154 0.0000

At most 1* 0.743769 121.3207 95.75366 0.0003

At most 2* 0.726134 81.83213 69.81889 0.0041

At most 3 0.606424 44.27376 47.85613 0.1044

At most 4 0.353306 17.23181 29.79707 0.6229

At most 5 0.138639 4.591243 15.49471 0.8506

At most 6 0.009036 0.263243 3.841466 0.6079

Trace test indicates 3 cointegrating eqn(s) at the 0.05

*denotes rejection of the hypothesis at the 0.05 level
**Mackinnon-Haug-Michelis (1999) p-values
Unrestricted Cointegration Rank Test (Maximum Eigenvalue)

0.05
Hypothesized Number of CE(s) Eigenvalue Trace Statistics Prob.**
Critical Value

None* 0.858577 56.72398 46.23142 0.0027

At most 1 0.743769 39.48859 40.07757 0.0581

At most 2* 0.726134 37.55838 33.87687 0.0173

At most 3 0.606424 27.04194 27.58434 0.0585

At most 4 0.353306 12.64057 21.13162 0.4858

At most 5 0.138639 4.327999 14.26460 0.8232

At most 6 0.009036 0.263243 3.841466 0.6079

Max-eigenvalue test indicates 1 cointegrating eqn (s) at the 0.05 level

*denotes rejection of the hypothesis at the 0.05 level
**Mackinnon-Haug-Michelis (1999) p-values
410 Impact of Climate Change and Economic Activity on Philippine Agriculture: A Cointegration and Causality Analysis

Table 3. Granger Causality

RELATIONSHIP
VARIABLES
UNILATERAL BILATERAL

Value Added of Agriculture Employment in Agriculture

Temperature Value Added of Agriculture

La Niña Value Added of Agriculture

Land Area Employment in Agriculture

Temperature Employment in Agriculture

Land Area Number of Typhoons that hit Landfall

Number of Typhoons that hit Landfall El Niño

El Niño La Niña

While the variables of the model are stationary at first Agricultural productivity to rainfall at 10% level of
difference, there is no guarantee that the regression results significance [23]. Conversely, data from 1976 to 2010 in
are not spurious. A more powerful test is needed for the Ghana, exhibited the cointegration and Granger causality
multivariate model employed in this paper. Using models of inter-annual yields of the crops have been
Johansen’s Cointegeration Test, the number of influenced by the total amounts of rainfall but changes in
cointegrating relationships is presented in Table 2. temperature were stationary, and were suspected to have
The maximum Eigenvalue and trace test reveal, minimal effect on crop yields [2].
respectively, 1 and 3 cointegrating equations. This Granger causality results of this paper show that value
indicates that the regression result is valid and that the added in agriculture exhibits a bilateral relationship with
aforementioned predictor variables and that Philippine employment in agriculture, which means that employment
agricultural production have a long-term equilibrium in agriculture impacts Philippine agriculture and vice versa.
relationship; consistent with the empirical evidence on the Factors of climate change – the La Niña and El Niño
long-run relationship between inter-annual changes in phenomenon, have an influence on each other. Conversely,
temperature and rainfall and different crop yields in temperature and La Niña exert a unilateral relationship
Northern Ghana [2]. with value added of agriculture; land area as well as
Since there exists cointegration among the variables of temperature affect employment in agriculture; and number
the model, testing for the direction of causal links can now of typhoons that hit landfall has a unilateral relationship
be carried using the Granger Causality test procedure. The with El Niño.
results are summarized in Table 3.
A similar study conducted in India made use of the Analysis of Regression Results and Diagnostic Tests
granger causality method that showed that water depletion
has a big impact on Agriculture production of crops. The An econometric model was created using time series
results of the Granger causality tests of the model show that data on the variables covered to validate the hypothesized
there is unidirectional causality between Agriculture relationships specified in the preceding chapter of this
production and Rainfall and that Rainfall Granger causes paper and the formulated hypotheses by employing
Agriculture production. The results provide a convincing applicable diagnostic tests. The following is a presentation
indication of a unidirectional causality consecutively from of the model's final results:
 Universal Journal of Agricultural Research 10(4): 405-416, 2022 411

lnVAA = 134.66 + .02 (lnEA) + .03(lnLA) - 7.71 (lnTYPH) - .08 (lnTEMP) - .02(D1) - .001(D2)
(0.01) (2.24) (0.05) (-0.06) (-3.75) (-2.73) (-0.91)
2
R = .99 F-stat = 406.83 D.W. = 1.98 s.e.e = 0.02
n = 31
Accordingly, the growing changes in temperature and production by 0.08%. Similarly, the incidence of El Nino,
precipitation as a result of climate change would lead to the other things equally, leads to a decline in agricultural
shift of production seasons, changes in pest and disease output by 0.02%; which is in conformity with a study in
forms, and would adjust the set of possible produce Australia that a moderate increase in both temperature and
affecting agricultural production, prices, incomes, rainfall and a moderate increase in temperature and a
profitability, opportunity cost of planting, and ultimately, decline in rainfall would affect agricultural productivity. It
livelihoods and lives [17, 36]. exhibited that a moderate increase in both temperature and
It was also found out the consequences of climate rainfall would yield increase by 2 to 9 per cent, and under
change on economic activity using temperature and market the second scenario – a moderate increase in temperature
prices of farms in the United States, which led to and a decline in rainfall are projected to increase yield by
established damage aggregation using the “bottom-up” 7 to 16 per cent [35]. The other predictors are to be
approach that would control diverse aspects of interpreted in similar fashion albeit they are not statistically
fundamental institutions. Accordingly, projections on significant.
climate change and the damaging effect on economic Comparably, research done in the Philippines between
activities remain high [27]. 2001 and 2009 that looked at the relationship between
Outcomes of the econometric model of this paper show rainfall and agricultural production revealed that
that Agricultural Expenditure and Agricultural Credit are above-average amounts of rainfall during the wet season
no longer included although they were part of the original are unfavorable for agricultural production, while
model. This will be explained shortly. Only three (3)
increased precipitation is beneficial during the dry season
factors reflected statistically significant effects on
[6]. In the same way, a study in Batac, Ilocos Norte on
agricultural production in the Philippines based on their
rainfall and temperature variability using annual data from
respective t-ratios: Agricultural Employment (EA),
1976 to 2010 showed that the consequences brought about
temperature (TEMP), and La Nina (D1). The rest of the
by climate change hamper agricultural productivity in the
predictor variables exert no statistically significant effects
province [13].
on agricultural production in the Philippines although their
In the same way, the impact of climate change in
respective signs are consistent with theoretical
Norway, as a result of changes in temperature and
expectations similar to the significant factors. Although
precipitation, on agricultural productivity over the period
taken collectively, all the predictor variables exert a
1958–2001 found out that in 18 per cent of cases,
significant effect on agricultural production.
The results are consistent with studies in Economic temperature positively impacts yield and in 20 per cent of
Community of West African States exhibiting the cases, precipitation has a negative impact on crop yields
influence of climate variables such as rainfall and [38]. In the same way, higher temperature would
temperature, and economic variables like capital and labor substantially affect crop yields and livestock, consequently
affects agricultural output [29]. In the same manner, it was farm incomes and food security [3].
found out that climate change in Nigeria deters the growth In China, it was likewise ascertained that the major
of the technology and manufacturing sector decreasing rice-producing regions experienced great yield losses
labor and capital productivity consequently hampering because of the occurrence of more frequent drought and it
growth and sustainable development [12, 34]. has threatened food security as well [24]. In Java, it was
The predictive power of the model is in fact very good also predicted that droughts have significantly affected rice
with an R2 of 99%. Literally, only 1% of variation in the production – small buildups in the global heat resulted to
dependent variable could be accounted for by factors, lower production on highly dependent rain areas [39].
which are excluded from the model. The deviation of the actual values vs fitted values
Other things equally, a 1% rise in agricultural estimated by the model is clustered around zero as seen in
employment gives rise to a 0.2% increase in agricultural Figure 1. This means that there is very little difference
production. On the other hand, a 1% rise in temperature, between the actual data series on agricultural production
other things equally, leads to a decline in agricultural and the estimates taken from the regression model.
412 Impact of Climate Change and Economic Activity on Philippine Agriculture: A Cointegration and Causality Analysis

13.8

13.6

13.4

13.2
.04
13.0
.02
12.8
.00

-.02

-.04

-.06
82 84 86 88 90 92 94 96 98 00 02 04 06 08 10

Residual Actual Fitted

Figure 1. Actual, Fitted and Residual Graph

13.8
Forecast: LNGVAF
13.7 Actual: LNGVA
13.6 Forecast sample: 1980 2014
Adjusted sample: 1982 2010
13.5 Included observations: 29
13.4 Root Mean Squared Error 0.020874
Mean Absolute Error 0.016609
13.3 Mean Abs. Percent Error 0.124715
13.2 Theil Inequality Coefficient 0.000783
Bias Proportion 0.000000
13.1 Variance Proportion 0.023256
13.0
Covariance Proportion 0.976744

12.9
82 84 86 88 90 92 94 96 98 00 02 04 06 08 10

LNGVAF ± 2 S.E.

Figure 2. Post-forecast Graph

Table 4. Variance Inflation Factor (VIF)

Independent Variables Code VIF Without AGRIEXP Without AGRICREDIT

Agricultural Expenditure AE 20.85178 - -
Agricultural Credit AC 8.786774 8.769968 -
Employment in Agriculture EA 1.406201 1.394538 1.379177
Land Area in Agriculture LA 9.253080 2.687609 2.676111
Carbon Emissions CE 11.29887 6.661678 2.111242
Precipitation PREC 3.433603 3.377799 3.082328
Number of Typhoons TYPH 2.137635 1.544439 1.514490
Temperature TEMP 2.148692 2.134577 1.569857
 Universal Journal of Agricultural Research 10(4): 405-416, 2022 413

Table 5. Final Diagnostic Test Results

DIAGNOSTIC TESTS T-STATISTICS CONCLUSION

Durbin-Watson Test Autocorrelation 1.98 Autocorrelation does not exist

LM Test Autocorrelation 0.95 Autocorrelation does not exist

Jarque Bera Normality of Errors 0.87 Error Terms are normally distributed

ARCH Heterokedsticity 0.24 Variables are homoskedastic

WHITE Heterokedsticity 0.61 Variables are homoskedastic

Chow Breakpoint Test Structural Stability Test 0.61 Model is structurally stable

Ramsey Reset Test Specification Error Test 0.09 Specification error does not exist

Forecast Performance of Model results, thus reducing the number of predictor variables
from nine (9) down to six (6). The final diagnostic tests
To find out further the predictive power of the model –
results are shown on Table 5.
ex-post forecast was applied using actual data series of the
The aforementioned summary results have ruled out the
variables from 2010-2014. Using the same equation of the
relevance of any diagnostic test breaches and show that the
econometric model, at +/- 2 standard error of forecasts,
model adequately satisfied the statistical and economic
which gives a 95% confidence interval that forecast will
requirements. The final regression result is exhibited in
fall within the interval, as shown in the resulting forecast.
Table 6, which now shows greater detail.
The predictive power of the model of this paper may be The Durbin-Watson statistics at 1.98 is approximately
gleaned from Figure 2. equal to 2.0 indicating that evidence of serial correlation
can be ruled out. The Breusch-Godfrey LM test was also
Post-Estimation Diagnostic Tests employed to corroborate the elimination of autocorrelation.
The p-value of the F-statistic at 0.95 percent exceeds
To ensure that data used in the study is not spurious, a the .05 percent level of significance suggesting that
post-estimation diagnostic test was carried out as shown in autocorrelation is no longer present.
Table 4. With a sample size of 29 observations, which may not be
Results suggest that the model adequately satisfied the large enough for a model using time series data, a test of
economic and statistical criteria and has ruled out breaches the normality of the regression residuals is vital to ensure
of pertinent diagnostic tests. that tests of significance of parameters are valid. Test
Table 4 demonstrates that as a measure of severe shows that the error terms are normally distributed given a
multicollinearity when testing for variance inflation factor p-value of 0.645 for the Jarque-Bera statistic.
(VIF). Agricultural Expenditure, Agricultural Credit, Land Test of heteroskedasticity for a regression model using
Area allocated to agriculture, and Carbon Emissions have time series data was implemented using the autoregressive
all exceeded 10. Agricultural Expenditure was also conditional heteroskedasticity (ARCH). Given a p-value of
excluded using sequential variable removal, having the 0.25 for the F-statistic, which exceeds 5 percent level of
highest VIF of 20.85. The regression model was rerun and significance, the regression residuals are deemed
tested again for VIF. With Agricultural Credit registering homoskedastic. Another heteroskedasticity test using
the highest VIF result of 8.77, this variable was also White procedure was used in the model, which yielded a
excluded from the model. Severe multicollinearity is no p-value of 0.61. This exceeds 5 percent level of
longer present in the remaining explanatory variables. significance suggesting that the residuals are indeed
Therefore, the problem on multicollinearity has been homoskedastic. A test of structural stability of the
resolved by deleting two independent variables. regression parameters was also carried out using the Chow
After eliminating multicollinearity, the model was Breakpoint Test. Since the p-values of the relevant F-ratio,
re-estimated and subjected to a new round of diagnostic chi-square, and Wald statistics all exceed 0.05 percent
tests. It indicated that the results are evidenced with level of significance, evidence of structural instability in
structurally unstable parameters, which will render the the model is ruled out. The model therefore is suitable for
model unsuitable for policy formulation and forecasting. use in policy formulation and forecasting. A serious
Moreover, there is also evidence of specification error in problem would still be possible if specification error exists
the model, which is quite serious. in the model. With a p-value of 0.09 on the F-statistic
As a remedial measure, Carbon Emission and which exceeds 0.05 percent level of significance, the
Precipitation, both indicators of climate change, had to be presence of specification error in the model is also ruled
removed from the previous model based on their VIF out. The model therefore is considered correctly specified.
414 Impact of Climate Change and Economic Activity on Philippine Agriculture: A Cointegration and Causality Analysis

Table 6. Final Regression Result

Dependent Variable: LNVAA

Method: Least Squares
Sample (adjusted): 1882 2010
Included Observations: 29 after adjustments
Convergence achieved after 180 iterartions

Variable Coefficient St. Error t-Statistics Prob.

C 134.6596 11462.11 0.011748 0.9907

LNEA 0.018877 0.008440 2.236685 0.0369

LNLA 0.029260 0.571585 0.051191 0.9597

TYPH -7.71E-05 0.001230 -0.062680 0.9506

TEMP -0.076168 0.020330 -3.746556 0.0013

LN -0.023224 0.008507 -2.732409 0.0128

EN 0.006348 0.006970 -0.910757 0.3733

AR (1) 1.157740 0.2345696 4.935034 0.0001

AR (2) -0.157926 0.236418 -0.667995 0.5118

R-squared 0.993892
Mean dependent var 13.32356
Adjusted R-squared 0.991449
S.D. dependent var 0.234350
S.E. of regression 0.021670
Alkaike info criterion -4.576628
Sum squared resid 0.009392
Schwarz criterion -4.152295
Log likelihood 75.36111
Hannan-Quinn criiter. -4.443733
F-statistics 406.8268
Durbin-Watson stat 1.982843
Prob (F-statistics) 0.000000

Evidence from both theory and practice illuminates how in agriculture to improve resilience, competence and
climate change has a negative impact on agricultural sustainability of the agriculture sector.
productivity, climate change significantly affects In order to attain agricultural productivity, food security,
agricultural productivity and impacts Philippine adaptation, and climate change mitigation, the Philippine
agriculture directly and indirectly. As demonstrated by the government should address elements that affect
findings of this paper, recognizing the risk posed by agricultural production; which are: Employment in
climate change and the relationship between each variable agriculture, and the incidence of the La Niña phenomenon
will help implement policy measures, increase the and Temperature, as variables of climate change.
resilience of Philippine agriculture to weather shocks, and
forecast long-term economic implications of climate
change.

REFERENCES
4. Conclusions [1] Abidoye, B., Odusola, A., “Climate Change and Economic
Growth in Africa: An Econometric Analysis”, Journal of
This study was carried out to gather data on the threats African Economies, vol. 24, no. 2, pp. 277-301, 2015. DOI:
10.1093/jae/eju033
that climate change poses to Philippine agriculture.
Regression analysis shows that the model is consistent with [2] Amikuzuno, J., Donkoh, S.A., “Climate Variability and
predictions made by many studies and journals cited in the Yield of Major Staple Food Crops in Northern Ghana”,
survey of related literature. African Crop Science Journal, vol. 20, s2, pp. 349 - 360.,
2012. https://bit.ly/3yY8fnV
The study shows that there is a long-term relationship
between value added in Agriculture in the Philippines and [3] Battisti, D.S., Naylor, R.L., “Historical Warnings of Future
the different economic and climatic variables. The Food Insecurity with Unprecedented Seasonal Heat”,
Science, vol. 323, no. 5911, pp. 240-4, 2009. DOI:
Philippines as basically an agricultural country, should 10.1126/science.1164363
assure the availability of vital inputs in agriculture, major
government expenditure should be redirected toward R&D [4] Beck, U., “How Climate Change Might Save the World:
 Universal Journal of Agricultural Research 10(4): 405-416, 2022 415

Metamorphosis”, Harvard Design Magazine, no. 39, 2013. Contribution of Working Group I to the Fourth Assessment
https://bit.ly/3NZdZ56 Report of the Intergovernmental Panel on Climate Change.
[Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.
[5] Cline, W., “Global warming and agriculture impact B. Averyt, M. Tignor and H. L. Miller (eds.)]. Cambridge
estimates by country”, Washington, DC: Center for Global University Press, Cambridge, United Kingdom and New
Development, no. 4037, 2007. https://bit.ly/3PaJH0m York, NY, USA, pp. 996, 2007. https://bit.ly/3Ixx9xZ
[6] Crost, B., Duquennois, C., Felter, J., Rees, D., “Climate [21] IPCC,” Workshop Report of the Intergovernmental Panel on
Change, Agricultural Production and Civil Conflict: Climate Change Workshop on Impacts of Ocean
Evidence from the Philippines”, IZA Discussion Paper, no. Acidification on Marine Biology and Ecosystems”, [C. B.
8965, 2015. DOI: 10.2139/ssrn.2598885 Field, V. Barros, T. F. Stocker, D. Qin, K.J. Mach, G.-K.
Plattner, M. D. Mastrandrea, M. Tignor and K. L. Ebi (eds.)].
[7] Department of Agriculture Annual Report, “Enabling
IPCC Working Group II Technical Support Unit, Carnegie
Communities, Expanding Opportunities”, PRDP, 2013.
Institution, Stanford, CA, USA, pp. 164, 2011.
https://bit.ly/3NXYKJJ
https://bit.ly/3O1Ac2x
[8] Department of Agriculture Annual Report, “The PNoy
[22] Johansen, S., “Estimation and Hypothesis Testing of
Administration’s Accomplishments in Small-Farmer and
Cointegration Vectors in Gaussian Vector Autoregressive
Fisherfolk Credit”, 2014. https://bit.ly/3c0JdeV
Models”, Econometrica. vol. 59, no. 6, pp. 1551-1580, 1991.
[9] Engle, R., Granger, C. W. J., “Co-Integration and Error DOI: 10.2307/2938278
Correction: Representation, Estimation, and Testing”, The
[23] Kamble, G., Honrao, P., “Impact of Environmental Change
Econometric Society, vol. 55, no. 2, pp. 251-276, 1987. DOI:
on Agriculture Production and Groundwater Depletion:
10.2307/1913236
Adaptation Strategies in Response to Farmers in
[10] Farnham, P. G., “Economics for Managers 3rd edition”, Maharashtra”, Journal of Economics and Sustainable
Pearson Publishing, 2014. https://bit.ly/3yuHyWI Development, vol. 6 no. 5, 2015. https://bit.ly/3NYpmui

[11] Fischer, G., Shah, M. M., van Velthuizen, H.T., “Climate [24] Li, R. and Geng, S., “Impacts of Climate Change on
change and Agricultural Vulnerability”, IIASA, 2002. Agriculture and Adaptive Strategies in China”, Journal of
http://pure.iiasa.ac.at/6670 Integrative Agriculture, vol. 12, issue 8, pp. 1402-140, 2013.
DOI: 10.1016/S2095-3119(13)60552-3
[12] Foye, O. V., Benjamin, O. O. (2021). “Climate change,
technology and manufacturing sector growth in oil-rich [25] Liu, J., “Stationarity of a Markov-Switching GARCH
Nigeria”, International Journal of Sustainable Economy., vol. Model”, Journal of Financial Econometrics, vol. 4, issue 4,
13, no. 3, pp. 236-260, 2021. DOI: pp. 573-593, 2006. DOI: 10.1093/jjfinec/nbl004
10.1504/IJSE.2021.116643 [26] Nicholson, W., Snyder, C. “Microeconomic Theory Basic
[13] Galacgac, E.S., Acosta, C.G, Alibuyog, N.R., “Variability of Principles and Extensions Tenth Edition”, Thomson
Rainfall and Temperature in Ilocos Norte, Philippines”, South-Western, 2008. https://bit.ly/3c5wLKS
MMSU Science and Technology Journal, vol. 3, no. 2, 2013. [27] Nordhaus, W.D., “Managing the Global Commons: The
https://bit.ly/3RrnQ6H Economics of Climate Change”, Cambridge, MA: MIT
Press, 1994. https://bit.ly/3IyI8ac
[14] Gornall, J., Betts, R., Burke, E., Clark, R., Camp, J., Willett,
K., Wiltshire, A., “Implications of climate change for [28] Nordhaus, W.D., “Estimates of the Social Cost of Carbon:
agricultural productivity in the early twenty-first century”, Concepts and Results from the DICE-2013R Model and
Phil. Trans. R. Soc. B, no. 365, pp. 2973-2989, 2010. DOI: Alternative Approaches”, Journal of the Association of
10.1098/rstb.2010.0158 Environmental and Resource Economists, vol. 1, no. ½, pp.
273-312, 2014. DOI: 10.1086/676035
[15] Gujarati, D.N., Porter, D.C., “Basic Econometrics”,
McGraw-Hill Irwin, 2009. https://bit.ly/3yxg2b0 [29] Ogbuabor, J.E., Orji, A., Manasseh, C., Anthony-Orji, O.I.,
“Climate Change and Agricultural Output in the ECOWAS
[16] Hilario, F., De Guzman, R., Ortega, D., Hayman, P., Region”, International Journal of Sustainable Economy, vol.
Alexander, B., “El Nino Oscillation in the Philippines: 12, no. 4, pp. 403-414. DOI: 10.1504/IJSE.2020.112304
Impacts, Forecasts, and Risk Management”, Philippine
Journal of Development, vol. XXXVI, no. 1, 2009. [30] Philippine Statistics Authority, “Selected Statistics on
https://bit.ly/3yVFt7y Agriculture “, 2015. https://bit.ly/2KvedjV
[17] Howden, M., Jones, R., “Risk assessment of climate change [31] Potsdam Institute for Climate Impact Research and Climate
impacts on Australia's wheat industry”, 4th International Analytics, “Turn Down the Heat: Climate Extremes,
Crop Science Congress Brisbane, Australia, 2004. Regional Impacts, and the Case for Resilience”, The World
https://bit.ly/3RkU1EZ Bank Group, 2013. https://bit.ly/3nSYa5o

[18] Hurduzeu, G., Kevorchian, C., Gavrilescu, C., Hurduzeu, R., [32] Prakash, G., Taylor, A., "Measuring Market Integration: A
“Hazards and risks in the Romanian agriculture due to Model of Arbitrage with an Econometric Application to the
climate changes”, Procedia Economics and Finance, vol. 8, Gold Standard, 1879-1913," NBER Working Papers 6073,
pp. 346–52, 2014. DOI: 10.1016/S2212-5671(14)00100-2 National Bureau of Economic Research, Inc. 1997. DOI:
10.3386/w6073
[19] Impact Forecasting, “Annual Global Climate and
Catastrophe Report”, 2013. https://aon.io/3ItaBOZ [33] Rudinas, J.S., Godilano, E.C., Ilaga, A.G, “Implementing
Climate Smart Agriculture Ridge – River - Reef: The
[20] IPCC, “Climate Change 2007: The Physical Science Basis”, Philippine Adaptation and Mitigation Initiative for
416 Impact of Climate Change and Economic Activity on Philippine Agriculture: A Cointegration and Causality Analysis

Agriculture”, Paper presented to the ASEAN-FAO-GIZ 10.1016/S0167-8809(02)00093-2

Regional Expert Forum on Climate Change: Agriculture and
Food Security in ASEAN. 2–3 May. Nai Lert Hotel, [37] Thamo, T., Adda, D., Kragt, M.E., Kingwell, R.S., Pannell,
Bangkok, 2013. https://bit.ly/3z0o6mc D.J., Robertson, M.J., “Climate change reduces the
mitigation obtainable from sequestration in an Australian
[34] Sam, Toong Hai, Zainab Lawal Gwadabe, Wong Chee Hoo, farming system”, Australian Journal of Agricultural and
Whee Yen Wong, Tan Seng Teck, Alex Hou Hong Ng, Resource Economics, vol. 63, Issue 4, pp. 841-865, 2019.
Asokan Vasudevan, Alvin Liau Chee How (2021). A DOI: 10.1111/1467-8489.12330
Review of Climate Change Impacts on Agribusiness Value
Chain in Nigeria. Universal Journal of Agricultural Research, [38] Torvanger, A.M., Twena, M., Romstad, B., “Climate
9(4), 119 - 126. DOI: 10.13189/ujar.2021.090403. Change Impacts on Agricultural Productivity in Norway”,
Draft Report. Center for International Climate and
[35] Salim, R., Islam, N., “Exploring the impact of R&D and Environmental Research—Oslo (CICERO) Working Paper
climate change on agricultural productivity growth: the case No. 2004:10, Blindern, Norway, 2004.https://bit.ly/3P24bZ
of Western Australia”, Australian Journal of Agricultural F
Resource Economics, vol. 54, Issue 4, pp. 561–582, 2010.
DOI: 10.1111/j.1467-8489.2010. 00514.x [39] Wassmann, R., Jagadish, S.V.K., Sumfleth, K., Pathak, H.,
Howell, G. Ismail, A., Serraj, R., Redonaa, E., Singh, R.K.,
[36] Tao, F., Yokozawa, M., Hayashi, Y., Lin, E., “Future Heur, S., “Regional Vulnerability of Climate Change
climate change, the agricultural water cycle, and agricultural Impacts on Asian Rice Production and Scope for
production in China”, Agriculture, Ecosystems and Adaptation”, Advances in Agronomy, vol. 102., pp. 91-133,
Environment, vol. 95, Issue 1-3, pp. 203-215, 2003. DOI: 2009. DOI: 10.1016/S0065-2113(09)01003-7