You are on page 1of 21

FORECASTING THE PRICE OF

CORN IN THE PHILIPPINES

BILLY JULIUS M. GESTIADA


December 14, 2017
MOTIVATION
Corn as the second most important crop
Limited studies
Imported goods
The hardships of corn farmers
INTRODUCTION
Information about corn
Corn farmers in the United States
US corn exports as earnings
BACKGROUND OF THE STUDY
Halonen (2016): supply and demand analysis forecasting
model
Bowman and Husain (2004): 14 commodities using
judgment-based, historical price-based, and commodity
futures-based
Jha and Sinha (2013): neural network model
OBJECTIVES OF THE STUDY
To provide a forecast on the price of corn in the Philippines

1) To describe the trend of the price of corn in the


Philippines from 1960 to 2016
2) To analyze the differences between the predictive data
and actual data
3) To investigate the performances of ARIMA and AR
models in forecasting the price of corn in the Philippines
SCOPE AND LIMITATIONS
80 provinces/cities in the Philippines including Metro
Manila
PSA is the only source of data
ARIMA and AR models were used
Data were from years 1960 to 2016, annually
CONCEPTUAL FRAMEWORK
METHODOLOGY
Research design
(a) historical research
(b) evaluative research
Source of Data: PSA, and Power and Intals World Bank
Comparative Studies (1990)
Two estimation models:
(a) Autoregressive Integrated Moving Average
(ARIMA)
(b) Autoregressive (AR)
METHODOLOGY
Predicted Value of FPRICE = + FPRICEt-1

FPRICE = Farmgate Price of Corn in the Philippines (in PhP/kg.)


t = Time
= Constant term, average change over time
METHODOLOGY
Predicted Value of FPRICE = FPRICEt1 + ut

= Phi coefficient (should not be less than 1)


u = Random error at period t
RESULTS AND DISCUSSION
45.00%
Education 40.00%
40.55%
36.29%

Higher education is more


35.00%
30.41%
30.00% 28.69%

invested in by women 25.00%

(PSA, 2015; Polachek, 2003; 20.00%


15.77%
18.99%

13.50%
Orbeta, 2002) 15.00%

10.00%
12.73%

Education is a determinant of 5.00% 2.53%


0.54%

income (OLS) 0.00%

Education signals necessary


skills for employment (Probit)
MALE FEMALE
RESULTS AND DISCUSSION
The results of the OLS regressions showed that:
MODEL 1 MODEL 2 MODEL 3 MODEL 4 MODEL 5
VARIABLES
Male Female Male Female Male Female Male Female Male Female

ELEM 0.230** 0.195 0.239** 0.181 0.159** 0.145 0.120** 0.122 0.053* 0.017

SECOND 0.551** 0.720** 0.568** 0.683** 0.388** 0.464** 0.318** 0.426** 0.216** 0.260*

COLLEGE 1.205** 1.562** 1.226** 1.525** 0.805** 0.930** 0.730** 0.883** 0.593** 0.678**

POSTC 1.707** 2.064** 1.718** 2.030** 1.093** 1.282** 1.010** 1.231** 0.872** 0.994**
RESULTS AND DISCUSSION
The results of the probit regressions showed that:
Highly-educated individuals have better chances of employment in
high-skilled and high-paying jobs (i.e. GOVT & PROF); whereas
elementary and secondary graduates have better chances at low-
and-medium-skilled occupations (i.e. SERVWORK, AGRIWORK,
TRADE_OPER, LABOR)
Women are often employed in professional occupations (33.6%), as
opposed to men (7.3%)
Men with less investment in education are often employed in
AGRIWORK, TRADE_OPER, and LABOR
RESULTS AND DISCUSSION
Years of Experience
LOGINC is insensitive to varying
years of experience; thus
explaining the narrow male-female
differential in terms of EXPER
Men earn more compared to
women due to the latters broken
labor force participation
(Polachek & Xiang, 2009)
Years of experience is often
insignificant in determining
employment and income
RESULTS AND DISCUSSION
Occupation VARIABLES
MODEL 5
Male Female
Individuals from high-skilled GOVT 0.528** 0.600**
occupations have higher returns than
PROF 0.758** 0.882**
others (i.e. GOVT, PROF)
TECH 0.502** 0.314*
For women, TECH, SERVWORK, and
CLERK 0.186** 0.434**
AGRIWORK are less significant, since
their investments are often not related SERVWORK 0.183** 0.237*

to the skill requirements of these AGRIWORK 0.192** 0.404*

occupations which determines income TRADE_OPER 0.212** 0.447**

(Javied, 2009)
SPECIAL 0.844** 0.053
RESULTS AND DISCUSSION
Industry
MODEL 5
VARIABLE
Although men earn lower returns in
Male Female
the agriculture sector than the
service sector, the latter was viewed
AGR -0.272** -0.058
as insignificant in determining
income
Women in the service industry earn SERVICE -0.009 0.171
higher returns than those from the
agriculture industry; however both
industries were insignificant in terms
of income
RESULTS AND DISCUSSION
Health
00% 86.54% 90.30% Majority of individuals have access to
00% 69.83%

00%
59.38% safe water and electricity, which could
00% indicate that these are cheaper today
00%

.00%
Safe Water Electricity MODEL 5
VARIABLE
MALE FEMALE
Male Female

The accessibility to safe water


WATER 0.234** 0.316**
and electricity, as measures of
health, were significant in income ELEC 0.273** 0.408**
RESULTS AND DISCUSSION
Human Capital Model vs
Job Signaling Model MODEL
F-TEST

Male Female
Using the F-test, all of the five
356.31** 118.49**
models for both the male and MODEL 1

female derived p-values that MODEL 2 335.21** 89.93**

are less than the 5% level of


117.71** 17.95**
significance MODEL 3

The human capital model was MODEL 4 99.64** 16.64**

accepted as a better explanation


of the Philippine labor market MODEL 5 74.86** 12.72**
RESULTS AND DISCUSSION
BLINDER-OAXACA
MALE-FEMALE
Male-Female Income VARIABLES
DECOMPOSITION (1973)
INCOME
DIFFERENTIAL
Endowment Remuneration
Differential CONSTANT
Effect Effect
0.5124 0.5124
ELEM 0.0137 -0.0004 0.0133
In total, men earn 8.59 SECOND 0.0377 -0.0310 0.0067
COLLEGE -0.1721 -0.0553 -0.2274
percentage points higher than the POSTC -0.0202 -0.0056 -0.0258

income of women EXPER


GOVT
-0.0027
-0.1444
0.0738
-0.0294
0.0710
-0.1737
PROF 0.0137 -0.0417 -0.0280
Women are still significantly TECH 0.1312 0.0547 0.1859

discriminated in returns that is CLERK


SERVWORK
-0.0197
-0.0019
-0.0457
-0.0113
-0.0654
-0.0131

24.25 percentage points lower AGRIWORK


TRADE_OPER
0.0026
0.0649
-0.0035
-0.0235
-0.0009
0.0414

than mens income SPECIAL


AGR
-0.0024
-0.0577
0.0074
-0.0170
0.0050
-0.0747

Men are more discriminated in SERVICE 0.0005 -0.1414 -0.1409


MALE-FEMALE
endowments INCOME
DIFFERENTIAL
-0.1567 0.2425 0.0859
CONCLUSION
Both the ARIMA and AR models have good overall
performance given their F-statistic values, and their
variables have significant p-statistics values as well
The AR model is the best model to forecast the corn
prices in the Philippines because:
- RMSE value is lower than ARIMA models
- MAPE value is lower than ARIMA models
RECOMMENDATIONS
Additional independent variables to account for other
factors
Government support to the farmers of corn for the
preservation of the commodity itself
The corn prices of neighboring ASEAN countries must be
included as well for a more comprehensive study
Corn prices can also be forecasted semi-annually,
quarterly, or monthly

You might also like