The Impact of Regional Fund on Unemployment

This research attempts to find the impact of the government budget on unemployment in West Java. It is conducted from 2006 until 2017. The data used here is secondary data from Badan Pusat Statistik (BPS). It consists of a general allocation fund (DAU), specific allocation fund (DAK), profit-share of tax and non-tax fund (Taxnon) as independent variables. Meanwhile, we set unemployment as the dependent variable. The method of analysis is Panel Regression with Fixed Effect. Because we firstly face awkward result when analyzing model by including all independent variables, then we try to estimate every single independent variable on the dependent variable. Therefore we have four models to dig the problem. Based on the analysis result, it is found that two independent variables have a negative and significant impact on the dependent variable. Meanwhile, profit-share of tax and non-tax fund (Taxnon) does not show significant impact, but only negative sign. From this finding, we can still say that unemployment can be on the wane due to various budget policies of a nation which directed to the regional development. The bigger the agenda of development, the bigger the decreasing unemployment rate. It is because everybody can be absorbed into employment.


INTRODUCTION
West Java Province is one of the provinces that occupies a prominent position, due to its location adjacent to the Capital of Jakarta.
However, the problem of unemployment in the region is quite alarming. The unemployment rate in West Java relatively high, more than 1.7 million people in 2015 (BPS Jawa Barat, 2017), compared to Central Java amounted to 0.86 million (BPS Jawa Tengah, 2015).
It is also higher if we compared to East Java. The unemployment in East Java is 906 thousand people (Kominfo Jawa Timur, 2017). Interestingly, West Java Province is a relatively important region in the presidential election process each period. The victory of all presidential candidates in Indonesia, whenever the electoral process carried out, is almost always supported by the voters of West Java that are large enough in number. In addition, nearly all analysts stated that the potential of natural and human resources possessed by West Java was quite large.
The main problem is because the amount of unemployment in West Java Province tends to be relatively large if we compared to other provinces Java island. Lots of allegations have emerged if the West Java province completely has been far from the development process in Indonesia, and only utilized its potential voice when facing general elections.
Among the instruments to reduce unemployment and increase development are the various types of budget programs: General Allocation Funds (DAU) and Special Allocation Funds (DAK).
Officially, the definition of DAU is explained in detail on the Ministry of Finance website (Kementerian Keuangan, 2016b) as follows: "The General Allocation Fund (DAU) is one of the transfers of government funds to regional governments sourced from APBN revenues, which are allocated with the aim of equitable distribution of financial capacity between regions to fund regional needs in the context of implementing decentralization." From the explanation above, it is clear that the implementation of the DAU program is aimed to improving economic conditions in the region. The focus is the life of people outside of big cities can prosper, and reach a level of prosperity. The government also run the DAU budget with DAK. It is explained in the Ministry of Finance website (Kementerian Keuangan, 2016a) the details about DAK budget as follows: "Special Allocation Funds (DAK) are funds sourced from APBN revenues allocated to certain regions to help to funding special activities which are regional affairs and according to national priorities." It is found that the General Allocation Fund (DAU) and poverty had a negative relationship (Istimal, 2012). Regression analysis in the research proves that when the DAU increases, the poverty rate in Tangerang city falls, although the value considered relatively small, according to the researcher. However, this remains proof that proper development planning can improve the welfare of the community.
On the other hand, we need to know that the sectors receiving the most Special Allocation Funds (DAK) are education, health, and road infrastructure. One research examined the use of DAK in the three sectors and found a uniformity of policies that did not provide opportunities for the possibility of different procedures for the conditions of each region in Indonesia (Usman, Mawardi, Poesoro, Suryahadi, & Sampford, 2008). This phenomenon is a problem. Because the DAK is intended to solve inter-regional inequality in terms of public services, we should apply policy adjustments to different conditions and situations. In practice, regional governments are only the passive recipients of these aid funds. The lack of initiative is a crucial problem so that the researchers suggest an idea of a new paradigm by breaking down authority through decentralization of allocation, coordination, and monitoring the use of funds for the targeted region.
The other findings (Setiyawati & Hamzah, 2007) are somewhat different from the of researchers in general. They found an interesting fact that PAD had a positive impact on economic growth, but DAU had a negative effect. The DAK and development spending are found not significant on economic growth. But economic growth was found to have a substantial impact on reducing poverty and unemployment. The subject of research consists of 29 districts and nine cities in the East Java Province.
Several researchers have succeeded in presenting empirical facts with the case of the city of Manado that the General Allocation Fund (DAU) and direct expenditure had a significant negative effect on the poverty rate (Paseki, Naukoko, & Wauran, 2014). Even their findings prove the direct and indirect influence of these two variables on poverty. The indirect impact of the DAU and direct spending on poverty rate are channeled through economic growth. The more the DAU and direct spending, the higher the economic growth; poverty rate decreases and welfare improves.
It is found empirical evidence that Revenue Sharing Funds (DBH) and DAU had a negative effect on poverty in Bali Province (Ismail & Hakim, 2014). But interestingly, DAK actually has a positive impact on the number of poor people. The population was found not to affect the amount of poor people. Meanwhile, the education and health proved to have a negative influence on the number of poor people. These things are indeed in line with the reality that proper education and health can reduce poverty amid society.
A study examined PAD, DAU, and DAK on economic growth and poverty in 33 provinces covering 441 districts/cities in Indonesia (Prasetyo & Aida, 2017). The results prove that PAD and DAU have a significant impact on increasing economic growth, and reducing poverty, whereas DAK was found to have no significant influence on economic growth or poverty.
On the contrary to the research before, it is found an empirical result that DAU had a significant adverse effect on economic growth (Astria, 2014). Meanwhile, capital expenditure has a significant positive impact on economic growth. The findings also show that both the DAU and capital expenditure have a value of elasticity that is greater than one. This finding means that the change in the two variables is one percent, which will make economic growth change greater than one percent.
Other study found evidence that PAD through regional spending was able to reduce poverty (Isramiwati, Rasuli, & Taufik, 2017). Meanwhile, DAU through regional expenditure does not have a significant impact on reducing poverty. The Revenue Sharing Funds (DBH) through regional expenditure have a significant effect on reducing poverty. The population through regional expenditure has a significant impact on reducing poverty.
A study presented empirically that DAK had several channels in reducing poverty in a region (Qomariyah, Suharno, & Priyarsono, 2016). Their results show that increasing DAK through the improvement of road and irrigation site can increase fiscal capacity, and reduce poverty in rural and urban areas.
A research found that in 129 districts/cities on the island of Borneo, PAD had a significant positive effect on the independence of a region (Tahar & Zakhiya, 2011). However, DAU had a significant negative impact on regional independence. But when the three variables were linked to economic growth, the results actually not significant. The authors eventually concluded that the absence of an impact on economic growth due to PAD, DAU, and independence was not directed to drive the economy of the community or the result was not too large on the economic activities of the residents.
In line with previous research, it is found evidence that the performance of earnings and Regional Original Income (PAD) had a positive effect on regional independence (Hadi, 2017). His research used the Government Financial Report (LKPD) in Central Java as the report for the study. From these findings, we can say that the better the earning performance and PAD of a region, the more independent the area.
Indeed, it needs to be stressed that the central government cannot control all regional matters. Regional budget planning is one of the advanced discourse in governance of Indonesia. It is concluded that economic growth, Regional Original Income (PAD), and the General Allocation Fund (DAU) had a significant positive effect on capital expenditure (Munir & Mahdar, 2016). Their research also showed the results that the implementation of regional autonomy allows each region to utilize their PAD and DAU for the benefit of development in their respective areas.
We need to recognize that research by several experts sometimes showed different results because they used different approaches or might use wrong calculation. One study found that PAD has been proved to have a significant positive impact on the economic growth of districts and cities in Central Java (Putri, 2015). However other variables such as DAU and inflation did not show significant signs of economic growth. These results are problem that should not be overstated, because it used panel models with only the common effect, without proposing more accurate test. Therefore this research cannot explain the phenomenon in more detail perspective with differences among the regions.
In another study, it is also found that DAU had a significant positive effect on employee expenditure (Samau, Rumate, & Londa, 2016). But the influence of the DAU was not apparent in capital expenditure in the Sangihe Islands Regency. Their research also explained that all this time the district government had relied too much on transferring funds from the central government, although its dependence had diminished over the past 3 years. This finding means that the empowerment of regional potential cannot be fully optimized.
It is concluded in a study that PAD, DAU and Profit Sharing Funds had a significant positive effect on regional expenditure (Putra & Dwirandra, 2015). Meanwhile, the DAK did not show a significant sign of regional spending. Their research findings also confirmed the absence of flypaper effects in Bali Province, which showed that local governments were not too dependent on transfer funds in making their expenditure formulations.
According to other researchers, the PAD and DAU had a significant positive effect on direct spending whereas DAK was found to have a significant negative impact (Hidayah & Setiyawati, 2014). Their research involved 34 districts/cities in Central Java Province. However, it seems that researchers do not look at the context of differences between districts/ cities that tend to be different, so that the method used only panel regression with a common effect, without the more detailed test.
We also found empirical evidence that DAU and DAK had a significant impact on regional expenditure (Gani & Kristanto, 2013). But the difference between them lied in the value of influence: DAU has a positive effect, while DAK is negative. The model used panel regression with a common effect involving 55 districts/cities on the island of Sumatra, from 2007 to 2010. The findings like this, namely DAK have a negative impact, warns us that sometimes economic research needs to consider various factors. It needs indeed to follow the procedure in the regression model, but also we must make sure regarding empirical facts among the societies.
In other research, it is found that only the DAK that had a significantly positive effect on regional government capital expenditure, but DAU did not has an impact (Machmud, 2013). The subjects of study are nine districts/cities in the North Sulawesi region, with a period from 2006 -2010. This study is the same as the other research that generated some anomalies. The researchers must be careful in concluding the results.
Interestingly, we also found the result that PAD and DAU did not affect direct government spending (Supadli, Gunawan, & Tamsah, 2018). Meanwhile DAK had a significant positive impact. Their research was conducted on local governments in West Sulawesi Province. This finding is different from the results of other expert study.
Other study found empirical evidence that the DAU, DAK, PAD had a significant positive effect on regional government capital expenditure (Juniawan & Suryantini, 2018). The study was conducted on eight districts and one city in Bali Province, for the 2014 -2016 budget year.
The study in the city of Manado showed us that both DAU and PAD had a significant and positive impact on increasing capital expenditure (Yawa & Runtu, 2015). The higher the DAU allocation, the higher the capital expenditure. Even with the increasing amount of PAD produced, it can grow the capacity of the capital city of Manado. This finding was supported by the other experts that found that DAU, DAK, and PAD had a significant positive impact on capital expenditure (Dewi & Suyanto, 2015) while the variable of economic growth did not have a considerable effect.
The budget that has been given to the regions has the aim to advance regional development. DAU, PAD, and DBH proved to have a significant positive effect on the human development index (Widarwanto & Yahya, 2014) meanwhile DAK and provincial financial assistance had no impact on the human development index. In other side, Basic Service Expenditure (BPD) which acts as a moderating variable can moderate the relationship between DAU, DAK, PAD, DBH, and BKP with the human development index in the districts/cities of the North Sumatra region. After inserted the moderating variable, DAK showed a significant positive effect on HDI whereas the other variables had no impact.
We also need to see the more significant potential of various development budget programs. The impact of tax revenue sharing and its role in reducing poverty sometimes were being questioned. Empirical findings proved that the increase in tax revenue sharing, agricultural expenditure, and industrial spending has a significant negative impact on local poverty (Lisna, Sinaga, Firdaus, & Sutomo, 2014). Their research also confirmed that increased fiscal capacity can reduce DAU, so that regional dependence on transfer funds can decrease. Their study involved 23 provinces in Indonesia from 2005 -2011.
It is clearly stated that various kinds of government funding programs such as DAU, DAK, PAD, and DBH played an essential role in growing the creative industry of a region (Zamzami & Hastuti, 2018). The research conducted in Jambi Province and proved all these variables had a significant positive impact on the processing industry. From this, we know that development funds are vital in supporting business progress so that the position is quite important to empower local potential.
Research in other countries can be used as a lesson. The exploration of the resilience of local fiscal policy in the state of Florida showed evidence that the spillover effect of changes in the distance in terms of miles in an area, as well as confidence in intergovernmental transfers, determined the financial conditions of various districts (Guo & Wang, 2017). This finding opens our view that spatial interactions in a region and the size of local savings have a strong relationship because it influences budget policies in the area. In the end, the community's development agenda was affected.
Indeed, it must be admitted that various economic programs to improve the standard of public life cannot be separated from multiple aspects. It is found that political decision roles often determine the number of budget allocations in European governments, both at regional and national levels (Bouvet & Dallérba, 2010). Therefore not only economic considerations but also the involvement of the political situation in various layers of government, determine the budget. Their findings confirmed how political factors play a role in translating the concept of development, which will be implemented in the European region. For this reason, as long as the government carries out the people's mandate correctly and adequately, this political aspect will not be a problem at all.
The regional financial agenda rolled out by the government to improve the local economy was essential to be seen as a form of state responsibility to mobilize various economic sectors to accommodate a large number of workers. The higher the development fund, the higher the expected development target. When the unemployment in an area is quite massive, we should question: To what extent large amounts of regional funds have been allocated.
According to BPS data, the number of unemployed people in West Java province as follows: From table 1 above, it can be seen that unemployment in West Java is quite massive. The authors consider the need for empirical test related to the role of government in efforts to tackle social problems that become obstacles to development.
Theoretically, in macroeconomic studies, there is short-term relationship of output deviation from its potential with an increase in unemployment. This theory is called Okun's Law (Ball, Leigh, & Loungani, 2017). In Okun's view, the higher unemployment indicates productivity in a region relatively low. This phenomenon is a problem that the stakeholders must overcome. Conversely, the higher economic growth signified unemployment rate declined. The role of the government in improving economic growth is very crucial.
Many economists have published research related to the role of taxes and government expenditures in encouraging increased economic standards in the community. Several experts conducted a test of government spending and its influence on economic growth in 182 countries in the world. The result they concluded that government spending had a positive impact on economic growth, without distinguishing the size and level of economic growth (Wu, Tang, & Lin, 2010). Interestingly, when the study conducted to differentiate the level of state income and the degree of corruption, it was found that in the case of developing countries, government expenditure did not show its significance. They concluded that corruption was a significant problem in development. From this finding, we can get an understanding that government conditions that are filled with corruption make the state budget ineffective and vulnerable to be abused.
One study concluded that trade taxation proved to have an inverse relationship with development, especially in countries with weak administrative capabilities (Adam, 2009). The benefits of tax in such states are only tool for the government to make money. The trading activities are faced with various difficulties and problems. The local government used to implement multiple reasons that manifested injustice in order their cash remained filled. The amount of tax that accompanied by poor government performance has a negative impact on economic development.
A study found an interesting fact that budget allocations in Bangladesh for some regions carried out in ways that were not elegant (Huque, 1992). In 12 local government regions, between 1986 and 1990, the utilization of local benefits was channeled with poor administrative capabilities, poor managerial knowledge, and abilities of the regional leaders, and failure to plan for efficient development. As a solution to overcome this problem, the author recommended a general improvement in the administrative field, by training local leaders and people's representatives in each region to expand their capabilities and knowledge. We cannot doubt that one reason for the high unemployment rate in Bangladesh is the bad budget management in various areas.
Based on the research of previous experts, we can say that the magnitude of unemployment in West Java, which according to BPS in millions, can be an indication of the lack of optimal productivity. It impressed that labor absorption appears weak, which denotes an inconsistency of the growth rate of Gross Regional Domestic Product (PDRB). In specific contexts, it is possible that economic growth in this province is supported more by sectors that are not able to absorb a lot of labor, or usually prioritize the capital (capital intensive). It could also be that economic growth only supported by several wealthy groups of people, while most of the population does not have any contribution except a little. Such possibilities can arise due to various facts from the reality that we face. All of the assumptions cannot be doubted because of the natural and human resources in West Java very large. If the great potential not maximally empowered, then certainly there is governance error.
Especially in recent times, the implementation of regional autonomy and fiscal decentralization are underway. Regions in Indonesia for example, now have autonomous rights to manage their own finances, without the need to wait for orders from the capital Jakarta. The empirical research in various countries has proven that the implementation of fiscal decentralization and government expenditure is in line with the improvement of people's living standards (Hessami, 2010). When the degree of autonomy of the local government increases, people's welfare increases too. It is a sign that the better lives of citizens are encouraged by an excellent public servant system.
It is found a quite exciting fact about fiscal. Some economists concluded in their research that the implementation of fiscal decentralization has indeed resulted in a better quality of government system, but on the other hand, it also contributes to higher income disparities in various regions of the country with weak governmental systems (Kyriacou, Muinelo-Gallo, & Roca-Sagalés, 2015). It indicates the importance of the role of transparent and accountable government administration. The Special Allocation Fund, and Profit-Share Fund on Unemployment message is apparent: the quality of public administration to increase people's welfare is very important. Not just a good development agenda, but it needs to be supported by the performance of a trustworthy institution.
The importance of this research is because we are trying to find out the truth of the empirical impact of the implementation of various development budgets by regional governments. Especially West Java as a case example in this research, we suppose there is a lot of confusion that although the province is close to the center of government in Indonesia, but the socio-economic empowerment of its people seems not optimal. We try to answer these negative issues with a scientific approach, through a test of their public planning budgets.
The results of this research are expected to be useful, both in terms of scientific and practical matters; the detail as follows: first, this research is able to enrich our discourse in regional finance, especially in the area of planning and allocation of DAU, DAK and the tax and non-tax revenue sharing. This research has advantages over previous other studies, namely in focus on the direct effect of the regional budget on unemployment. Other experts usually study regional budgets directed to its ability to increase regional economic growth through PDRB or to decrease poverty. Besides, this study also opens our view that the determination of variables in a study requires relevant consideration. It is not only from an econometric perspective but also economic theory and facts in general; second, the findings in this study can add to the reference in consideration of stakeholders concerning the management of regional budget allocations, in overcoming various social problems in the community specifically unemployment.

METHOD
The research methodology consists of an explanation of the data sources, collection, and analysis methods used to test the proposed hypothesis.
This research uses quantitative data from BPS (Central Bureau of Statistics) and Ministry of Finance. It consists of unemployment, DAU (General Allocation Fund), DAK (Special Allocation Fund), and tax and non-tax revenue sharing funds, in several areas in West Java province (BPS (Badan Pusat Statistik)). This data are provided free of charge by the government of West Java and Ministry of Finance, as a public reading material that can be accessed by anyone. The researchers and observers can learn it without any obstacles. The method of data collection in this study was obtained from the budget report provided by the BPS of West Java province and Ministry of Finance for some regions. The study was conducted from 2006 to 2017.
In this study, the data of unemployed used as the dependent variable or the affected variable. While data of Tax and Non-Tax Revenue Sharing, General Allocation Funds (DAU), and Special Allocation Funds (DAK), all three are independent or influencing variables. All of these were obtained from the West Java Central Statistics Agency (BPS) report. The analytical method used in this study is panel regression with fixed effects. The selection of panel regression is considered as the most feasible tool because this research involves a combination of data time-series and crosssection. To make the analysis more manageable and capturing growth rate, we convert data into natural logarithms.
We should make several tests before conducting a panel analysis. Firstly we make the selection of feasibility between common effects and fixed effects. Secondly, we make the choice between random effects and fixed effects. This selection indicates how to obtain the accuracy of the model that will be used, to find the accuracy of the results. Panel regression is a combination of time series and cross-section data. With this method, we are expected to be able to find the results of an analysis that is closer to the truth instead of using only one data instrument (Baltagi, 2005).
In this study, the author began by making a panel regression analysis model with 3 independent variables (Tax and Non-Tax Revenue Sharing, General Allocation Funds (DAU), and Special Allocation Funds (DAK) and one dependent variable (unemployment). All data for each selected region are included as part of the analysis. Nonetheless, we found some problems with the results. Therefore, we suggest further analysis with the same model (panel regression), but the different variables slightly.
In the second, third, and fourth models, the author determines the panel regression model for each independent variable with the unemployment as dependent variabel. We believe that to see the influence of each of the independent variables, we do not have to rely on entering all of them in one model. We can make an equation model for one independent variable and one dependent variable. Therefore, the estimation is made into four models.
Model 1 for this study as follows: Description: Unemit = Number of Unemployment in period t, for region i Taxnonit = Tax and Non-Tax Revenue Sharing in period t, for region i DAUit = General Allocation Funds (DAU) in period t, for region i DAKit = Special Allocation Funds (DAK) in period t, for region i = error term period t, for region i Model 2 formed as follows: Description: Unemit = Number of Unemployment in period t, for region i Taxnonit = Tax and Non-Tax Revenue Sharing in period t, for region i = error term period t, for region i Model 3 formed as follows: = error term period t, for region i

RESULTS AND DISCUSSION
From the empirical findings in model 1, we get the following results: From the equation in model 1, we can find out that only one independent variable (DAK) having a significant effect on the number of unemployed people in West Java. This result means that the government's budget policy aimed to increasing development in West Java is not optimal. We cannot reject the estimation because some eligibility tests have been run to support the result.
The results of the test of model 1 above are supported by the estimation of the feasibility test between the common effect and fixed effect. Result of calculation can be seen in appendix 1.1. Based on the calculation, the F-stat value is 60.405186. In table 2, we know that this study has a denominator of 228 and a numerator of 4. Referring to references, the value of the F-table at the degrees of error α = 1% and α = 5% are 3.48 and 2.45 respectively (Widarjono, 2009).
Thus, the F-stat value in the table is higher than the F-table, so that the alternative hypothesis is accepted, namely fixed effect is chosen as the proper model.
Similarly, it is also supported by the result of the feasibility test between the common effect and random effect. The value of Breusch-Pagan show significance. The test can be seen in appendix 1.2.
The model also supported by the estimation between fixed effect and random effects (Hausman Test).
Based on table in appendix 1.3, it is proven that the chi-square of the calculation result is 75.487322. While the value of the chi-squares table at the degree of freedom of 3 at α = 1% and α = 5% are respectively 11.34 and 7.81 (Widarjono, 2009). It can be concluded that the value of the chi-squares in the table 4 is higher than the standard chi-squares. Thus, the fixed effect is preferred over random effect. From here we have got a good estimation result and met the econometric standard. All tests have been proven its feasibility test.
But the findings on the first model can surprise researchers, especially those who have only preliminary experiences in economic analysis. Although the regression analysis meets the standard requirements, there is a big problem we must face. In the first model, we find that all independent variables do not affect unemployment. The variation of the dependent variable cannot be influenced by the independent variables. Why this happens.
The main reason for this, we guess, that each variable has characteristics that cannot be embedded with other data. From the three independent variables, for example, DAU, and DAK have different specifications. Although each policy in its determination involves central finance. But the specs of it are not the same.
If we believe in the results in model 1, we can be suggested that all government policies will be considered not to be significant in reducing unemployment. We could say it useless and might be assumed that the government had wasted money without empirical results. It could also be, in a more extreme context, we guess the current government conducted corruption of public funds. The conclusion of this analysis is problematic. Here we need to re-question not only the regression analysis method, that must be valid, but it needs to refer back to the economic theory that has been used as a foothold in the research.
We should see the phenomena in the real context. This action cannot be ignored. We also need to refer to the references that have been convinced: the government budget plays an essential role in creating welfare among the communities. Therefore, the results in model 1, although it followed the regression method, but violated the standard rules in the study of economics. An anomaly in the data often occurs. We cannot discuss anything further from the calculation of model 1, because all independent variables do not affect unemployment.
This result reminds us that the econometrics model requires a solid and solid theory, not only statistically correct but also in line with an economic perspective. The econometrics is only a tool to detect an economic phenomenon that occurs in society, not the final goal. The real purpose of an econometric model is to uncover that occurs in the real world. If in the end the econometric search contradicts the economic theory, then we need to go back to the original idea of what econometric model was made for. We must not be too fixated on the econometric model, but forget the essential substance of a real economic study: get clarity of an economic phenomenon in society.
We go to the next model. The result of the model 2 as follows: From the calculation above, it is known that the tax and non-tax profit sharing does not have a significant negative effect on unemployment. This situation is evidenced by the value of t-statistic -1.4626 which is lower than the standard t-table at 1 percent and 5 percent namely 2.617 and 1.980, respectively (Widarjono, 2009). The negative sign (-) in the table above shows the negative effect which is in line with the purpose of the program. R 2 value: 0,90, indicates that the variation of the dependent variable can be explained by the independent variable by 91 percent. While the rest explained other things outside the model. This finding clearly reminds us regarding problems that we must resolve in the real context.
The test the feasibility of the model has confirmed that the fixed effect is more appropriate than the common effect. Based on table in appendix 2.1, it is known that the F-stat value is 83.480807. According to the result, this test has a denominator 228 and numerator as much as 2. The F-stat value is higher than the Ftable at 1 percent and 5 percent degree of freedom, which show value 4.79 and 3.07 consecutively (Widarjono, 2009). Therefore the common effect is rejected.
Similarly, it is also supported by the result of the feasibility test between the common effect and random effect. The value of Breusch-Pagan show significance. The test can be seen in appendix 2.2.
Analysis of the feasibility test between fixed effect and random effect also proved an exciting result. Based on table in appendix 2.3, it Special Allocation Fund, and Profit-Share Fund on Unemployment is known that the value of chi-squares in the calculation results is 12.779062. While the critical value of the chi-squares distribution table with a degree of freedom 1 percent and 5 percent respectively are 6.63 and 3.84 (Widarjono, 2009). Then it can be concluded that the value of the calculated chi-squares is higher. So the fixed effect is more feasible than the random effect.
Based on calculations in model 2, we can be sure that government policies for tax and non-tax revenue sharing basically play a significant role in reducing unemployment, but our estimation cannot accept it. We believe that government agenda conducted, both central and regional, throughout the world, has one single purpose: to eradicate unemployment. None of them made policies in vain to mitigate the size of the unemployment rate. The findings in model 1 and 2, namely the absence of influence from tax and non-tax profits, must be questionable.
There are many reasons why this happens. In a country or region that has an high level of corruption, a tax budget of any magnitude will not be able to improve people's welfare. In Indonesia, the level of corruption rate is indeed high. But so far, the process of law enforcement has continued and ensnared the perpetrators of corruption firmly. We can witness by ourselves the reality of how the eradication of corruption takes place. Many tax mafias have been caught and sentenced.
Next, we go to the analysis in model 3. The results as follow: This result means that if there is an increase in the DAU of 1 percent, unemployment can fall by 0.09 percent. This value is reinforced by t-stat which is 1.881 is higher than t-table at 10 percent degree, namely 1.658 (Widarjono, 2009). The negative sign (-) in the table above shows the negative effect of variable X on Y. the value of R 2 is 0.90 indicating that the variation in the dependent variable can be explained by an independent variable of 90 percent. While the rest explained other things outside the model. The findings in this model also refute the panel regression results in model 1, which shows that there is no influence from DAU on unemployment.
Testing for model accuracy proves that fixed effects are more appropriate than common effects. According to the estimation in table appendix 3.1, the value of F count with the numerator 2 and the denominator 228 is 67.089214. This value is higher than the F-table both at 1 percent error rate and 5 percent for 4.79 and 3.07 respectively (Widarjono, 2009). Therefore, the common effect is rejected.
It is also supported by the result of the feasibility test between the common effect and random effect. The value of Breusch-Pagan show significance, lower than 1 percent. The test can be seen in appendix 3.2.
The comparison test to choose between random effects and fixed effects shows that the first type is the most feasible. The results of the analysis can be seen in appendix 3.3.
Based on table, we get the value of the chisquare of 42.101098. The results of the calculation notified that it is higher than the value in the chisquares table at 1 percent and 5 percent, which are equal to 6.63 and 3.84 respectively (Widarjono, 2009). So that we choose the fixed effect proved to be more feasible than random effects. The results in model 3 provide information that DAU (allocation of general funds) has a significant positive impact in reducing unemployment. The government through this budget tries to offer various opportunities that can provide employment. All governments in any country or region in the world is willing unemployment declines. It is because unemployment is a burden of development. People's welfare will not be achieved if unemployment high. The failure to overcome unemployment is a sign of the collapse of the government to carry out its role in improving welfare. Governments everywhere will fight hard to solve the unemployment problem.
The last model, namely the fourth model, for the analysis in this study obtained as follows: From this, we can say if there is an increase in DAK of 1 percent, unemployment can fall by 0.06 percent. This value is strengthened by tcount which is 2,852 which is higher than ttable at 5 percent degree which is 1,980 (Widarjono, 2009). The negative sign (-) in the table above shows the negative effect of variable X on Y. The value of R2 of 0.90 indicates that the independent variable can explain the variation in the dependent variable at 90 percent while the rest explained other things outside the model. This finding refutes the results in model 1, which shows that there is no effect of DAK on unemployment.
The DAK budget has become one of the important pillars of improvement program of welfare. In order to erode unemployment, various methods have been carried out by the government. The agenda that the government implements will not be covered with any misuses, especially at this time, transparency in public finances can be easily questioned by various groups of people. Nowadays it is not difficult to find people who misuse the municipal budget in the area. The ease of access to information has opened opportunities for civil society to participate in monitoring the budget of development. To eradicate unemployment rate is one of the programs set as a priority development agenda. The reason, unemployment is the root of all kinds of crimes and various other evils that emerge among societies. When unemployment is rampant, greater social problems arise. Not only ordinary crimes, but other bad behaviors will circulate. The unemployment matter will lead to poverty and social instability that destroys the social order. The findings of model 4 also supported the measure of model feasibility. It finds evidence that fixed effects are more appropriate than common effects. The estimation table can be seen in table appendix 4.1.
The estimation shows the calculated F value of 110.015455. This value is higher than the F-table both at 1 percent error rate and 5 percent for 4.79 and 3.07 respectively (Widarjono, 2009). Therefore, the common effect is rejected.
It is also supported by the result of the feasibility test between the common effect and random effect. The value of Breusch-Pagan show significance. The test can be seen in appendix 4.2.
The comparison test between choosing a fixed effect with random effects shows evidence that the fixed effect is more feasible to choose. The estimated results are in appendix 4.3. Based on calculations, the value of the chi-square was found at 4.051768. The results of this calculation Special Allocation Fund, and Profit-Share Fund on Unemployment are higher than the value of the chi-squares table at 5 percent, which is 3.84 (Widarjono, 2009). Therefore we can still choose fixed effects, as a more feasible model than random effects.
The results of this study may not be very similar to several studies that have been done by other researchers. Generally, the researchers examine the relationship between DAU and DAK with economic growth and or regional government spending. There are also several studies that try to find the effect of these two variables on poverty. Our research attempts to explore what many economists have never done before, namely looking for the impact of budget allocations (DAU, DAK, tax and non-tax revenue sharing) on unemployment. We deliberately look for empirical facts, whether there is direct evidence of the influence of DAU, DAK, and tax and non-tax revenue sharing on unemployment.
Some studies such as conducted by Paseki, Naukoko, & Warran (2014) show the negative influence of DAU on poverty. If the assumption that poverty is close to unemployment, then our research has almost the same direction as their findings. Unemployment can be the root of the problem of poverty. People become poor due to the loss of income earned from their work. When their work is deprived, with the termination of employment, many people end up living in misery and not a few who are depressed then end their lives. So we conclude from their findings that our research though used different variables but still had the same spectrum.
Research conducted by Qomariyah, Suharno, & Priyarsono (2016) seems to lead to the same conclusion, namely DAK has a negative impact on poverty through several channels. Our research is quite different. Besides the various dependent variables, we also do not mention the existence of specific channels. The negative effect on unemployment has the same inlet as poverty: empowerment of people. But the process of achieving it can be very different.
Meanwhile, research by Munir & Mahdar (2016) shows the positive impact of DAU on capital expenditure. We know that the purpose of capital expenditure is to improve people's welfare. There is no capital expenditure agenda that aims to be in vain. Our research findings have a dependent variable which, although different, aims to find the impact of the DAU on unemployment. Unemployment is a problem that can be eradicated by increasing capital expenditure for community empowerment. The greater the resources we spend on them, the greater the achievements of empowerment: the decline in unemployment and achieving full employment-despite in fact this situation is complicated to achieve.
For some cases, the findings of this study implicitly contradict those obtained by other experts. For example the research conducted by Setiawati & Hamzah (2007) found a negative influence of the DAU on economic growth. We suspect, there are several reasons might generate it: first, the data is not smooth enough. It is strange, if the DAU has a negative impact on economic growth because the DAU funds are allocated to finance various development activities. Therefore according to theory, DAU funds will have a positive effect on economic growth. Second, we assume that the research conducted by them is faced with a bad situation. It could be, if we might assume, they were researching in an area where the government did not support development achievements. Back to the economic theory and it has been proven everywhere, DAU is one of the instruments to achieve the development agenda. Therefore we suspect there is something wrong, due to their findings contradict. Third, their research seems to be too fixated on the results of econometric estimation. Our research is actually, with the results presented in model 1, econometrically correct. There is absolutely no violation that can make the finding invalid. However, we do not believe in the result. Therefore, we created simulations with different models: we reestimate by allowing each independent variable to be directed to the dependent variable immediately one by one. We comment on their research that they need to review the data, by first refining it in order do not produce bias and or other anomalies.
Ismail & Hakim's research (2014) found different results between the influence of DAU and DAK on poverty. We see the phenomenon of different findings as one of the big problems that must be resolved by the policymakers. Their results indeed can be casuistic and cannot be generalized to the others. Many kinds of research produced by researchers prove the negative influence of both the DAU and DAK on various development problems, including poverty. The program and the regional financial agenda were made because the local government tried hard to mitigate social problems. We suspect the findings produced by Ismail & Hakim (2014) are somewhat similar to those conducted by Setiawati & Hamzah (2007). Their data need refinement, in order not to incur irrelevant result. Our research was also initially problematic because it showed anomaly when DAU, DAK, and Tax and Non-Tax Revenue Sharing did not affect unemployment. The purpose of the three financial agendas is clear to improve the living standard of the people, including reducing the unemployment rate. Various essential problems of the regional development process can be eradicated with a proportional budget and the right allocation. We should trace back the challenge of result and proposed a recalculation with an econometric model that was more in line with economic theory, and not just numerical calculations. Any number if we put into econometrics software will generate a result, even if it is entirely unrelated and no relevance at all; like for example we regress between the number of rats at home with our monthly income. Even though those two variables do not have a correlation and strange analysis of course, if the two variables are forced to be calculated by using software such as Eviews, STATA, JMulti, R or Gretl, the output will definitely appear, although all the results are meaningless nonsense.
After trying hard to grasp the phenomenon, we also still get into an anomaly: profit-share of tax and non-tax fund does not show significant impact. However, this finding can be reminder for regarding the problem of tax in our country, especially in West Java region. We should see this as "a hidden message" of development that must be resolved nicely.
Based on the findings in this study, we get an understanding that the magnitude of the financial budget in the West Java area clearly plays a role in reducing the unemployment rate. Of course what we hope is that the budget allocation directed to important posts in the community empowerment program. Indeed good budget allocation is one of the important aspects that local governments must play. This application is crucial, because it involves the lives of many people. The misuses of allocation with a poor administrative system will clearly have a negative impact on the program.
The corrupt system often arises from the government that is not sensitive to problems amid society, such as market failures. For example, when prices of basic needs increase, not everything can be left to the central government. The local government should be the one who knows the most problems in their area, because they have the most authority to find solutions at the first time. That is one of the important tasks of the regional government, in ensuring the fulfillment of the rights of all citizens. Special Allocation Fund, and Profit-Share Fund on Unemployment Conversely, if the government actively involved in activities beyond its role, the implications will be adverse for the people, for example, by raising local taxes without considering the economic conditions of the local society. The policy only adds to the local treasury, but the end is detrimental to many people. With higher taxes, many economic activities will be hampered. Even in the long run, the higher the size of the tax, the smaller the actual level of tax revenue. This is due to the decrease in economic performance among society. We can say roughly that there will be many businesses to go bankrupt because the taxes are too heavy.

CONCLUSION
This research has succeeded in showing through scientific instruments that the budget allocation in DAU and DAK has a negative effect on unemployment in West Java. Meanwhile, the distribution of tax and nontax returns cannot influence significant effect, but only show negative sign. This finding means that empirically, the government programs have a real impact on advancing welfare. The larger the budget program, the lower the unemployment rate. The estimation is evidence of the improvement of socioeconomic conditions of the community. The increase of unemployment means that there is a problem that must be resolved.
In this research also, the author proposes evidence that the use of precision regression models is a necessity to obtain robust analysis results. Based on the findings, it showed us that the selection of the correct research model is not always following facts and reality, even if it has fulfilled the element of feasibility. Therefore, in this research we present several models in order to be able to be considered as a proper model by other observers and researchers.
We must admit that initially our study found no significant effect of all the independent variables on the dependent variable. We doubt, and finally re-estimate it by proposing a different model. We tested the independent variables one by one on the dependent variable. The results prove that each independent variable has a significant effect on the dependent variable. We reject test resulted in the first model, although it met the standard of the econometrics.
Some of the results of other experts who display oddities and a lot of anomalies need to be criticized. We do not just take the results they found. For the number of studies that are contrary to general economic theory and reality, we propose as a comparison of the findings in our research. It could be that the results with the anomaly were due to conditions or extreme situations that led to strange outcomes. We can only guess it because economic studies basically cannot rely solely on the analysis of numbers alone.
Nevertheless, this research still has several weaknesses. The author recommends that future researchers examine the phenomenon of unemployment with a more complex perspective with different methodologies. It could also need to involve a longer observation time with the study area that not just in West Java. However, other researchers can make this research as a reference to achieve a more comprehensive understanding of the relationship between budget allocation policies and unemployment.