Selected Factors Affecting the Subsectors of the Philippine Agriculture: A Panel Regression Analysis

The study examined the effects of exports to Gross Domestic Product (GDP), employment, and production loans granted in the GDP agriculture of the Philippines. This paper employed descriptive and quantitative techniques to analyze the behavior of GDP, production loans granted (PLG), exports to GDP (EGDP), and employment (EMP) from 2005 to 2015 totalling 33 observations of agriculture, forestry, and fishery sectors. Specifically, panel data analysis was used to assess the effects of APLG, EMP, GDPt-1 and AEGDPt-1 in GDP. The fixed effect model corrected from autocorrelation and heteroscedasticity, a one percentage unit increase in the exports to GDP, on the average, leads to Php 774.96 increase in the GDP, other things equal; and a one-unit increase in employment, on the average, leads to PhP23.55 increase in the GDP, other things equal; a one peso increase in the production loans granted lagged by one period, on the average, leads to Php 0.4760 increase in the GDP, other things equal. Using the fixed effect model, all the explanatory variables such as, exports to GDP, employment and production loans granted lagged by one period exhibited significant effect on the GDP agriculture. Hence, the model is considered satisfactory from statistical perspective. The results from the fixed effect model were consistent with the priori expectations that exports to GDP, employment and production loans granted lagged by one period positively affect the GDP agriculture.


Introduction
Agriculture is the art and science of managing plant and animal growth for human use. It includes farming and all its branches such as soil cultivation or tillage, production, planting, and harvesting of all agriculture and horticulture products, rearing of animals and poultry, dairy and practices performed out by the farmer on a farm as an incident to or in conjunction with such farms. It also includes forestry and logging operations but excludes the manufacture or processing of farm products.
The Philippines is generally an agricultural country notwithstanding its attempt to make it an industrialized nation in 2000. According to the Philippine Department of Agriculture (DA), the Philippines is an agriculture country with a land area of 300,000 km2, 47 percent of it is comprised of agricultural land. The key agricultural lands are clustered around large urban areas with high population density. The agricultural sector of the country is divided into farming, fisheries, livestock/poultry, and forestry. The major agriculture crops are coconut, rice, bananas, corn, coffee, sugarcane, pineapple, mangoes, tobacco, and abaca. Based on the World Bank data, the contribution of agriculture in Philippine GDP from 1960 to 2016 has an average value of 21 percent with the lowest of 10 percent in 2016 and the highest of 31 percent in 1974.
The country is also the second largest archipelagic state and one of the largest fishing countries in the world. In 2012, with a production of 3 million tons of fish, mollusks, crustaceans and other marine animals, the Philippines topped the global major fish producing countries.
Aquaculture contributed to total fish production by 790 900 tons, or 25.4 percent. Most of its output is locally consumed with a consumption of 33 kg per caput of fish in 2011. Furthermore, in 2012, the Philippines is the third largest farmed seaweed producer with a production of 1.8 million tons. An estimated 1.5 million people were employed by the fisheries and aquaculture industry nationwide in 2010, with fisheries accounting for over 1 million. The fishing industry is contributing an estimated 1.24% (priced at 223 billion pesos) to the current prices of the Gross Domestic Product (GDP), respectively in 2020. Based on Food and Agriculture Organization 2014 report, the fish and fishery products exports were valued at USD 1.2 billion in 2013. Tuna was the top commodity for export, followed by prawns and shrimps. Forests are also very important during climate change. The planting of trees also reduces atmospheric carbon dioxide and reduces the possibility of climate-related disasters such as storms and drainage.
The government has acknowledged the decreasing contribution of the agricultural sector in the country's GDP, from fourteen percent (14%) GDP share in 2000 to eight percent (8%) GDP share in 2018. This is primarily due to extreme natural disaster, infestation of the pests and lack of high-quality of crops at the end of the harvest. The maximization of land potential is prevented because of limited diversification of agricultural product such as rice, corn, and sugarcane. Issues such as lack of irrigation, limited access to financial institutions, outdated farm machineries and scarce post-harvest facilities, ageing farmers, among others have also been recognized.
With the government's efforts to address the problems of agriculture sector, the Philippine Development Plan 2017-2022 will secure the sustainability of its endeavor to enhance the productivity and continuously build the capacity of its stakeholders. Some of the development plan strategies are to improve productivity within ecological limit, increase agriculture-based enterprises, access to technology, financial inclusion, and ensure and protect land tenure security of agrarian reform beneficiaries.
This study is an attempt to analyze the behavior Philippine agriculture output along with selected exploratory or independent variables such as, exports to GDP, employment, and production loans granted across the different sectors of agriculture for the period of 2005 to 2015. Specifically, this study finds answers to the following research questions: 1. Is there a significant correlation among exports to GDP ratio, employment, production loans granted, and agricultural output? 2. Do production loans granted, exports to GDP ratio, employment significantly affect the agricultural output, when taken individually and collectively?
3. Is the effect of the explanatory variables mediated by the differences across agricultural subsectors?
4. Do the subsector differences significantly affect the agricultural output? 5. What is the appropriate panel regression model to be used as model for Agricultural Output?

Theoretical framework
Production Theory Basics. The main concept of production is to transform the resources into capital goods. Production, storage, shipping, and packaging can be included. Economists broadly describe production as something other than consumption economic activity. Any commercial activity other than the final purchase is considered a form of production. The production process is carried out in time and space. It is a concept of flow that determined by means of output rate per time. The production process covers three aspects: the quantity of the produced commodity, the quality of the produced commodity, and the temporal and spatial distribution of the produced commodity. Any activity which increases the similarity between the demand pattern for products and the quantity, forms and distribution of such goods on the market can be described as a process of production.
Economists refer to the resources or inputs used for the production process as production factors. Various potential inputs are typically grouped into four or five categories such as commodities (natural capital), labor (human capital), capital goods and land. The fifth category that is the managerial and entrepreneurial skills is added considering that it is a subcategory of labor services. Capital goods are those goods that have previously go through a production process. Other authors considered technology as factor of production process. Management can adjust all these factors in the long-term production. However, the short-term production usually consider at least one factor of production must be permanent. Fixed production factor is a factor whose amount cannot easily change such as large equipment, such as suitable production areas, and key management personnel. Variable production factors are factors whose level of use can be easily changed like consumption of agricultural products, transportation services and most raw materials. In the short-term production, the company's business scope determines how many products can be maximally produced. In the long-term production, there is no limitation in the scale. • P is total production (the monetary value of all goods produced in a year)

Cobb-Douglas Production
• L is labor input (the total number of person-hours worked in a year) • K is capital input (the monetary worth of all machinery, equipment, and buildings) • b is total factor productivity • α and β are the output elasticities of labor and capital, respectively. These values are constants dictated by technologies available.

Export-Led Growth Theory.
The export-led growth hypothesis (ELGH) assumes that increasing one's country exports results in increasing the growth of the entire economy. Exports can lead as a "growth engine" based on its advocates.
The relationship between growth and export is widely related to the probable effective externalities for the local economy resulting from participation in global markets, e.g. the reallocation of existing capital, economies of scale and other impacts of labor training. These mechanisms, however, are often used without any theoretical support or scientific evidence. A significant number of ELGH research has been carried out in developing countries (DCs) over the last 30 years. In fact, a new series of scientific studies on a number of different research lines, methodologies, time periods and countries were conducted during the 1990s (Smith, 2001).

Agriculture-based economic development is an agricultural-based economic
development strategy that involves a shift in technological, institutional and financial incentives change that will increase small farmers' productivity. Wiggins (2006) explained that in the process of economic growth, agricultural financial incentives play a dual role. First, it will provide more food and create many great jobs that are needed as well. The theory targets the methods by which underdeveloped countries can turn their domestic economic systems from a heavy focus on traditional livelihood farming to a more current and advanced farming practice through significant financial support to achieve an industrial breakthrough (Igyo, 2016).

The Importance of Exports and Labor Productivity in Agriculture
China is the world's most populated nation and also a country with numerous farmers.
According to 2005 World Bank data, forty-five percent of total employment is in agriculture. In signifies that the country is fully involved in economic globalization and its agriculture has become part of the globalization process.
The economic crisis will affect the U.S. agriculture directly and indirectly. Changes in U.S. economy will cause a direct effect. On the other hand, the intensity of crisis impacts foreign revenue and trade and global energy prices will result to indirect effects. The growth of the dollar for declining world GDP will significantly reduce import and agricultural work and reduce export and agricultural prices. The main element of uncertainty in the long term is the exchange rate of the dollar, especially in relation to the Chinese yuan. The dollar can still be measured or depreciated. In the latter case, the return of global growth combined with valuable dollars would create a strong foreign demand for United States agricultural products.
The importance of the link between sustainable development and agricultural exports is much depicted in growing economies. Agriculture serves as a significant driver for economic development mostly for developing countries. Countries like Thailand, Pakistan, China and Brazil experienced a high capital for production but gains low income because of the degradation of resources and low prices of small farmer. It is clear that fair trade export scheme can be an essential tool to make exports a benefit to local communities (Fan et.al, 2013).
According to Al Hamwi (2005), Syria sees the potential of export trade improve their GDP by means of raising the export earnings. Their product seeks to develop competitiveness in the world market. The government uses policies such as anti-export distortions and restrictions on imports in order to satisfy its agricultural products. The effect of these policies has been hampered before the state interference in price, trade, and foreign exchange controls, thus reducing exports of some commodities, banning other exports and export taxation. Now that their economic reform has been carried out, agricultural policies such as exemption of taxes in exported agricultural product, elimination of connection between the imports and exports, reduction of aerial transportation for exports, used machinery classifications, sorting and packing have been allowed to be imported, and an exchange rate has been unified.
According to Umar (2010), agriculture is Nigeria's largest economic activity. The estimated 70 percent of the total workforce is involved in agriculture. While the oil sector is primarily responsible for the growth and development of the Nigerian economy, agriculture is still a main source of economic sustainability. Nigeria accepts substantial amounts of natural rubber and cocoa exports. However, large export income variations raised concerns over the potential to grow and sustain the country's future.
Other policies like tariffs that maintain high domestic prices and high domestic support interrelate export subsides. These measures can result in overproduction or export subsidies on the world markets with adverse effects on non-subsidiary producer countries. These countries could impose high tariffs in order to protect their producers from cheap subsidized imports.
Therefore, a revision of internal policy and the reduction of import tariffs could go hand in hand with eliminating export subsidies. Developing countries would benefit from liberalizing their own markets to abolish export subsidies without costly changes (Peters, 2006).
According to Ellis (2008), the main dependence on the agriculture is still on weather.
Climate change in numerous parts of the world has already had a detrimental effect on agriculture due to growing weather conditions. The Food and Agriculture Organization (FAO) released a caution that a sudden rise in global average temperature of two to four degrees Celsius over pre-industrial levels is enough to diminish the crop yield of Africa and Western Asia by 15-35 percent, and 25-35 percent in the Middle East. The study concludes that there is no concrete, long-term solution for this problem that is published in the existing literature on agriculture, trade and climate change. The rising demand for energy-intensive food also increases the volume of agricultural commodities in international market. How the greenhouse gas emissions from processes like transport and packaging decreases related to trade must be continually emphasized. Governments and policymakers need to focus emphasizing attention to direct and indirect discharges from export-oriented farming, such as production, packaging, transport and storage methods.
Agricultural activities are seen lower for the past decade but still contribute to the economy since bulk of its labor force which compromise 37 percent comes from agriculture. The production of livestock is an important contributor to the gross domestic product in many developing countries. Agro-forestry practices mainly focusing on trees, forages and cattle are known as Silvopasture. Forages are deliberately introduced into wood production systems or trees are deliberately placed into a drilling system is the process of silvopasture systems.
Through the interactions between timber, drilling and livestock, wood, high quality drilling resources and effective livestock production can be produced at the same moment. Silvopasture can generate overall economic returns and create a sustainable system that delivers many benefits for the environment. These systems range from traditional silvopastoral to fodder cuts and systems of very high intensity. Macabasco (2004) surveyed the state of the Philippine Fruit Industry. It argued that fruits are major dollar earners and major sources of employment in the Philippine economy. Fruit production grew by 3.6 percent from 1998-2002, which was mainly driven by the 6.5 percent increase of banana production during the same time frame. Banana accounts for 60 percent of total fruit volume during the past five years (1999)(2000)(2001)(2002)(2003)(2004), while pineapple and mango account for 20 percent and 10 percent of fruit production, respectively. As such, these three crops account for 90 percent of the total fruit production of the country. With regard to trade, Macabasco stated that fresh and processed fruits reached over $600 million (FOB) yearly from 1998-2002. The country also exports papayas, jackfruit, lanzones and calamansi and the major destination for these fruits is Japan, accounting for over 40 percent of the total value of fresh/dried fruit exports in 2003. The other markets are China, U.S.A., South Korea, U.A.E., and Taiwan.
Lastly, Macabasco (2004) also analyzed the strengths, weaknesses, opportunities, and threats to the Philippine fruit industry. It argued that the fruit industry continues to suffer due to untapped potentials. The main constraints remain to be limited application of recommended production technologies and low productivity, inefficient marketing infrastructures, lack of longterm financing and largely unorganized producers. The main advantage is agro-climactic endowment and pockets of global excellence. The argument therefore shows the need for increased government assistance to be able to help exporters overcome the said constraints. As such, Macabasco concludes that the Philippine fruit industry must ensure the production of high quality products to remain competitive in the world markets. This high-quality production and competitiveness are based on three key factors: production, technology, and management.
The study of Habito et al. (2005) reviewed the trends in the Philippine agricultural sector's performance, and relate it to the policy environment within which the sector has operated through the years. It specifically examines the trends in production and productivity of the sector and how was it influence by the policies created. The results show that Philippine agricultural sector deteriorated over time compared to its neighboring countries. Akram et al. (2008) defined the constraints and proposed remedial steps that would allow efficient use of agricultural credit schemes. Most farmers reported that due to the need for collateral, they could not use credit. The rough hits were tenants and shareholders who do not own the property and were therefore unable to take advantage of credit. Another constraint was the high mark up from both formal and informal sources. The respondents' borrowing behavior was estimated using the logit model and the determinants of credit constraints were identified.

The Role of Financial Markets to Agriculture Output
Results showed that the transitory income, educational level, and forecast interest rate coefficients have significant impacts on borrowing behaviour. The expenditure on household consumption was determined positively and significantly by operating holdings and equipment value. The study showed a highly significant long-term crop area coefficient since the land is a fundamental input to the agricultural sector. The study also showed that the proportion of soil intensity increases the sector output. Scientific results show that direct credit has a positive relationship with agricultural output. Credit always helps purchase various inputs from this sector, so it has an indirect role to play.

Measure of Correlation
The Pearson Correlation Coefficient (r) is employed with the following formula to calculate the correlation between variables (Walpole et al. 2007, 434): The value of r close to one implies very strong correlation or degree of linear association between X and Y, while values close to zero indicate little or no correlation, ceteris paribus.

Measure of Regression
The research utilized three (3) panel regression techniques to analyze the behavior of GDP. The following model was utilized in the study:

Tests on Panel Regression Model: a. Stationarity Test
The Levin-Lin-Chu (LCC) Test procedurewhich assumes a common unit root process is used to test for stationarity of the data and the existence of panel unit root with null hypothesis that each time series contains a unit root against the alternative hypothesis that each time series is stationary (Kunst, Nell & Zimmermann 2011, 1).
Where the lag order p is permitted to vary across individuals (The null hypothesis is p = 0). The procedure work first by running augment Dickey-Fuller (ADF) for each cross-section on the equation (2): The second step is taken for the two auxiliary regressions (2): The residuals are then standardized with the following procedures (3): Finally, run the OLS pooled regression: If | |< , ̅ is said to be weakly (trend-) stationary, while if | |= then ̅ contains a unit root.

b. Pooled Least Squares
The estimation of the model depends on the assumption of the researcher who makes on the intercept, the slope coefficients and the error term (Gujarati 2004, 640). The following are the assumed Fixed Effect Model (FEM) and Random Effect Model (REM) that were applied in this study.

Fixed Effect (Cross-section) or Least Squares Dummy Variable Model (LSDV)
Employing the fixed effect model, which allows the slopes to vary across the classification of the sectors of agriculture, the differential slope dummies were added to the model as follows:

Random Effect (Cross-section) Model
The REM allows the slopes of the models as random variable with mean value of for the first model and , where i = 1, 2, 3 and j = 1, 2.

c. Test for individual Significance
To test for individual statistical significance of exogenous variables, the t-test is used with the following the formula: If the computed t-statistic exceeds the tabular t value, or critical region at α = 0.05, the null hypothesis that =0 is rejected. Alternatively, if the p-value of the computed value is less than the chosen level of significance, the null hypothesis is rejected. iv. Test for Overall Significance To test for the overall significance of the regression model, the F-test is used. It tests the overall explanatory power of the parameters using the following formula: If the value of the F-statistics exceeds the critical value of the distribution at 5 percent level of significance with k-1 and n-k degrees of freedom, it can be inferred that not all of the regression coefficients are not equal to zero and thus the regression estimate is significant and the model is to be considered valid (Gujarati, 2004). v. Measure of Multiple Determination to test how well the sample regression line fits the data, the coefficient of regression, R2, is employed with formulas as follows: However, the adjusted coefficient of determination, 2̅ ̅ ̅ ̅ , was used to penalize for the reduction in degrees of freedom as the number of injected independent variables in the regression increase.

d. Durbin Watson Test to Autocorrelation
To check for the presence of autocorrelation or the assumption that the error terms are not correlated, the Durbin-Watson statistic, d, is used with the following formula: The calculated value of d ranges between 0 and 4, with no autocorrelation when d is in the neighborhood of 2. If the computed d-statistics is greater than the upper limit of d (du) but less than 4-du (i.e., du < d < 4-du), there will be no evidence of positive or negative autocorrelation (Gujarati, 2004).

e. Test for Homoscedasticity
To determine the assumption of Homoscedasticity, that is, the error terms in the regression model is of equal variances, the study utilizes the Breusch-Pagan Test.

Findings and Discussions
The results from Pearson correlation tests indicate that the Gross Domestic Product (GDP) have significant and strong positive linear relationship with the all the variables production loans granted (PLG), exports to GDP ratio (EGDP), and employment (EMP). Given each p-value of 0.0000, which is below the five percent level of significance and thus, REJECT the null hypothesis that these independent variables have no significant correlation with the GDP. Panel unit root tests using Levin, Len and Chu Statistic (LLC) method were conducted on stacked series. The results in Table 6 showed that GDP and Production Loans Granted are nonstochastic at their levels. The Exports to GDP is non-stochastic at its first difference while the Employment is non-stochastic at second difference. The adjusted R-squared value of 0.9995 in the initial pooled model means that 99.95 percent of the variation in the GDP is explained by Exports to GDP (EGDP), employment (EMP), production loans granted (PLG) lagged by one period and GDP lagged by one period taking into account the number of explanatory variables used in the model. As expected, all explanatory variables have positive coefficients except for production loans granted (PLG) lagged by one period. The negative coefficient is in line with related literature of the effect of payments and costs of borrowing in the GDP. The null hypothesis for Exports to GDP, GDP lagged by one period and for production loans granted (PLG) lagged by one period variables was rejected at 0.0017, 0.0000, and 0.020 probability, respectively.
However, there is presence of positive autocorrelation. The DW statistic of 1.501 falls below 1.739 du < dw < 2.21 4-du. The second pooled model in table 4 (Table 4).  The model in Table 6   The Wald test has a 577.06 F-statistic and a probability of less than 5 percent, thus concluding the Fixed Effect model is the appropriate model than the Pooled Regression. In relation to the fixed effect model, pooled regression is a restrictive model in that it imposes a common intercept on all regions. Therefore, restricted F-test is used. The F-value is: The F-value of 31.28 exceeds the critical F-value of 2.93 at five (5) percent level is highly significant. Therefore, the restricted regressor is invalid. Specifically, fixed effect model is better than the pooled regression model.

Conclusion
In the final pooled regression model, the positive sign preceding the coefficients of exports to GDP, employment and GDP lagged by one period, indicate a positive relationship with the dependent variable GDP except for production loans granted which resulted a negative relationship. This relationship is in accordance with the theoretical framework and literature discussed in this study.
The results from the unit root test on panel data using the Levin, Lin and Chu procedure suggested that the variable GDP and production loans granted are stationary at level, the export to GDP ratio is stationary at first difference and the employment at its second difference. All of the explanatory variables have strong positive correlation with dependent variable, GDP.
Based on the fixed effect model corrected from autocorrelation and heteroscedasticity, a one percentage unit increase in the exports to GDP, on the average, leads to Php 774.96 increase in the GDP, other things equal. A one-unit increase in employment, on the average, leads to PhP23.55 increase in the GDP, other things equal. Lastly, one-peso increase in the production loans granted lagged by one period, on the average, leads to Php 0.4760 increase in the GDP, other things equal.
Using the fixed effect model, all the explanatory variables such as, exports to GDP, employment, GDP lagged by one period, and production loans granted lagged by one period exerted significant effect on the GDP agriculture. Hence, the model is considered satisfactory from statistical perspective.
The results from the fixed effect model were consistent with the prior expectations that exports to GDP, employment and production loans granted lagged by one period positively affect the GDP agriculture.
The result from the effects of exports to GDP, employment and production loans granted lagged by one period are positive, hence, the following recommendations are suggested: 1. Policy options such as reducing cost of borrowing to small farmer, effective crop insurance program with immediate claim, effective guarantee program for smallholder financing, and easing access to financial intermediaries for the farmers are essentials to increase the agricultural output.
2. The private sector plays an important role in agriculture. Increase responsible and productive investments in agriculture and enhance competence in the supply chain.
Expand entrepreneurship to create an employment and stable local agribusinesses.
3. Increase investment in forestry by creating market for forest services. Developing plantations that will facilitate the legal harvesting of forest products for household and industry consumption.
4. Locally mobilizing investment funds is very important in meeting at least part of the investment needs of the future. And to attract participation from the private sector, it is necessary to make available investment profiles of appropriate projects with relevant information and analysis.
5. As there is a potential for Philippine agriculture products and the global price becomes competitive, it is recommended for the government to support export activities for farmers. An increase in access to international markets, improved participation in global value chain and addressing prospects for a regional trade are relevant moves.

Acknowledgement/Any declaration
The author would like to acknowledge Dr. Gregoria Guce, Prof. Ricardo Dizon, Prof.
Melcah Monsura, Dr. Aileen Camba, and Prof. Jo-Dann Darong for making a critical review of the paper which resulted in its further improvement.