The Epirical Econoics Letters, 1(12): (Deceber 211) ISSN 1681 8997 Deterinants of Uneployent in the Philippines 1 Thirunauarasu Subraania Departent. of Southeast Asian Studies, University of Malaya, Kuala Lupur, Malaysia Eail: starasu@u.edu.y Ahad Zubaidi Baharushah Departent of Econoics, University Putra Malaysia, Malaysia Eail: baharushah@yahoo.co Abstract: This paper exaines the deterinants of uneployent in the Philippines using the Autoregressive Distributed Lag Method (ARDL). This study reveals that the Philippines labour aret ay be affected by a structural uneployent proble. Political instability ay also retard econoic growth and aggravate the uneployent proble in the Philippines. Thus, it is iportant for the Philippines staeholders to engage in active labour aret policies. At the sae tie, aintaining a stable governent is also crucial to the Philippines econoy. Keywords: Political change, uneployent, Philippines JEL Classification Nubers: E24 I. Introduction Deterinants of uneployent include job vacancy, econoic growth, foreign direct investent, exports and governent spending. Downes (1998) analyzed the uneployent issue in Trinidad and Tobago and concluded that econoic growth plays a doinant role in uneployent reduction. Other factors deeed iportant by Downes include increases in real wages and real loan, which adversely affects the uneployent rate especially in the long-run. Kooros (26) highlighted an inverse relationship of uneployent with Gross National Product (GNP), discount rate and inflation, but a positive relationship with the wage rate in the case of the United States. The ipact of active onetary policy on uneployent in the Philippines was analysed by Canlas (1997) and the author concluded that active onetary policy is not effective in 1 The authors wish to than the participants of International Networ for Econoic Research (INFER) Worshop entitled Changes in the Labour Maret and Challenges of the 21st Century held at University of Rovira I Virgili, Tortosa, Spain fro 1th -11th of April 28 for their constructive coents and suggestions.
The Epirical Econoics Letters, 1(12): (Deceber 211) 1248 uneployent reduction, as it only produces a teporary decline in uneployent and the downside will be the peranent inflation rate that ushers in. Trivedi and Baer (1985) concluded that a rise in uneployent in Australia is related to rise in real unit labour costs, lagged uneployent and real uneployent benefits (prior to 1975), and a decline in the rate of capacity utilization. Valentine (1993) asserted that the growth in real unit labour costs is the ain cause for an increase in uneployent in Australia. Christopoulos and Tsionas (22) suggested that a reduction of the governent sector is an additional channel through which eployent could grow faster. Barro (1991) and Scully (1989) opined that there is an inverse relationship between governent size and econoic growth. On the other hand, Abras (1999) highlighted a positive relationship between uneployent and governent size. Driffield and Taylor (2) in their analysis on foreign direct investent (FDI) and the labour aret found that FDI potentially increases wage inequality and the use of relatively ore silled labour in doestic firs. Ruane and Gorg (1999) showed that in the Republic of Ireland, FDI contributed to aggregate eployent. Seyf (2) in his study on the uneployent proble in the European Union (EU) concluded that the sae volue of FDI created ore jobs in Gerany copared to any other EU countries. The author also concluded that it is unliely that encouraging FDI would enable European countries to reduce uneployent substantially. Teo et al. (24) studied the uneployent-vacancy relationship for Singapore, Japan, Hong Kong and Taiwan and concluded that these countries are experiencing greater uneployent and low vacancies. According to the authors, the higher negative coefficient obtained for the vacancy rate in Singapore suggests that Singapore s labour aret is experiencing higher atching efficiency copared to Japan and Korea. Figure 1 below shows the econoic growth, uneployent and job vacancy rate in the Philippines fro 1974 to 23. The econoic growth was negative in the Philippines during the period of 1984-1985, 1991-1992 and 1998. Uneployent during these periods were also found to be high. In the 198s and 199s, the econoic growth in the Philippines had not exceeded 5 percent. In 2, two years after the 1997 Asian financial crisis, the econoic growth in the Philippines was 3.9 percent. Even with positive econoic growth, the uneployent rate reains high in the Philippines.
The Epirical Econoics Letters, 1(12): (Deceber 211) 1249 Figure 1: GDP Growth, Uneployent Rate and Job Vacancy Rate Relationship, the Philippines, 1974-23 15 1 5 Rate 1974 1976 1978 198 1982 1984 1986 1988 199 1992 1994 1996 1998 2 22-5 -1-15 Year ur vr eg When econoic growth was the lowest in 1985, uneployent rate increased. A severe recession in 1984-85 saw the econoy shrin by ore than 1 per cent, and perceptions of political instability during the Aquino adinistration further dapened econoic activity. 2 The econoic downturn of the early 198s was very daaging to the Philippines, which slid into a full-fledged depression because of the worsening ters of trade, a large accuulation of foreign debt that led to a stoppage of international credit, and the collapse of the anufacturing sector (Galenson, 1992). It is obvious that the Philippines econoy is siply not able to create enough jobs, causing the uneployent rate to reach doubledigit. The slow econoic growth in the Philippines coupled by political instability could be the potential causes. The ipact of political change on uneployent is rather neglected. Thus, this study intends to loo at the ipact of political change on uneployent in the Philippines. With this in ind, the objectives of this paper are organized as follows: a. to deterine whether there exists a long-run relationship aong uneployent, job vacancy, FDI, econoic growth, governent spending and inflation; b. to identify the ain acroeconoic variables that deterines uneployent in the Philippines; 2 http://en.wiipedia.org/wii/econoy_of_the_philippines accessed on 7th of February 27.
The Epirical Econoics Letters, 1(12): (Deceber 211) 125 c. to analyse the ipact of political change on uneployent. The organization of this paper is as follows. Section 2 briefly explains the data sources; Section 3 discusses the odel specification; Section 4 loo at the ethodology; Section 5 gives the epirical results and Section 6 concludes. 2. Data Secondary data will be utilized in this research. This data is obtained fro World Developent Indicator 26 CD-ROM, published by the World Ban. Data used in this study is annual data that covers the period fro 1974 to 23. Data on job vacancy rate for the Philippines was obtained fro Yearboo of Labor Statistics, published by the Departent of Labour and Eployent Philippines. 3. Model Specification Based on the uneployent odels developed by Phelps (1994), Layard and Nicell (1986), Assarson and Janson (1998) and Teo et al, (24), we posit the following specification to deterine the uneployent rate for the Philippines. UR t = β + β1vrt + β 2 EGt + β 3GOVt + β 4 FDI t + β 5INFt + β 6 DUM + ε t (1) where URt is the uneployent rate; VR t is the job vacancy rate ; EG t is the econoic growth; GOV t is the governent spending; FDI t is foreign direct investent; INF t is the inflation rate; and DUM is the duy variable. Variable DUM is represented by DUM 1 (which represents the 1997 Asian econoic crisis) and DUM 2 (which represents the 1986 Philippines Revolution). The expected signs for all coefficients are expected to be negative as per the theory. 4. Methodology Pesaran et al. (21) suggested the use of the Autoregressive Distributed Lag Model (ARDL) approach when the tie series under study coprise of a ixture of I() and I(1) and it is sall in saple size. The Unrestricted Error Correction Model (UECM) or error correction version of the ARDL approach can be written as follows: DURt 1 i t 1 2 i t i 3i t 4iDFDIt + i= i= i= i= = bo + b DVR + b DEG + b DGOV + b b5 iinft 1 + b6urt + b7vrt + b8 EGt + b9govt + b1fdit + b11inft + u1 t (2) i=
The Epirical Econoics Letters, 1(12): (Deceber 211) 1251 DURt = bo + b1 i DVRt 1 + b2 i DEG t i + b3i DGOVt + b4i DFDIt + i= i= i= i= b5 iinft 1 + b6urt + b7vrt + b8 EGt + b9govt + b1fdit + i= b11 INFt 1 + 12DUM1 + u2t α (3) DURt = bo + b1 idvrt 1 + b2 i DEG t i + b3i DGOVt + b4i DFDIt + i= i= i= i= b5 iinft 1 + b6urt + b7vrt + b8 EGt + b9govt + b1fdit + i= b11 INFt 1 + 12 DUM 2 + u3t γ (4) where D represent the first difference operator; represent the lag length selected based on the Aaie Inforation Criteria (AIC); u 1t, u 2t and u 3t are norally distributed residuals. The bounds test developed by Pesaran et Al. anipulates the F-statistic by iposing certain restrictions on the level lagged variables. The null and alternative hypotheses are as follows: H : b6 = b7 = b8 = b9 = b1 = b11 = (5) H 1 : b6 b7 b8 b9 b1 b11 (6) The null hypothesis above indicates that the long-run relationship is non-existent in the three odels. Vice versa, the alternative hypothesis, indicates that at least one of the b i is not equal to zero. The F-value test (Wald test) coputed after iposing the restrictions will be copared with critical bound values as proposed by Pesaran et al. The nullhypothesis of a non-existence of a long-run relationship can be rejected if the coputed F- value exceeds the upper critical bound. Vice versa, the null of no cointegration cannot be rejected if the coputed F-value falls below the lower bound. A conclusive inference cannot be ade if the F-value coputed falls between the upper and the lower bound. Equations 2, 3 and 4 above, augent the short-run as well as the long-run effects. The short-run effect is captured by the coefficients of the first-differenced variables. The long run coefficients can be obtained by ultiplying the coefficients of the one-lagged explanatory variables with a negative sign and subsequently dividing the values obtained with the coefficient of one-lagged dependent variable (Barsden, 1989; Tang, 22).
The Epirical Econoics Letters, 1(12): (Deceber 211) 1252 5. Epirical Results Table 1 suarizes the results of the unit root test, which eploys the Augented Dicey-Fuller (ADF) and Phillip-Perron (PP) tests for the Philippines. All series naely UR, VR, EG, GOV, FDI and INF are found to be a ixture of I() and I(1) for the Philippines. This gives us a good basis to eploy the ARDL cointegration approach. The unit root test is perfored in order to differentiate stationary and non-stationary series and at the sae tie to deterine whether the variables under investigation is I() or I(1) processes. Deterination of the order of integration is iportant in order to identify the use of appropriate cointegration techniques. The lag length selection for the ADF test is based on the AIC, whereas for the PP test, the Newey-West bandwith is allowed to autoatically choose the optial lag-length. Table 1: Unit Root Test Variables ADF PP C C&T C C&T Panel I: Level UR -1.93 -.525-1.527-3.262* VR -.16 -.72 -.263-1.77 FDI -2.756* -3.421* -2.673* -3.412* EG -3.418** -3.345** -2.938* -2.877 GOV -1.485-2.523-1.22-1.592 INF -4.595*** -4.395*** -4.59*** -4.767*** Panel II: 1 st Difference UR -4.87*** -4.767*** -8.128*** -7.977*** VR -6.179*** -6.433*** -5.414*** -8.117*** FDI -7.*** -6.899*** -11.726*** -12.357*** EG -4.24*** -4.223** -6.198*** -6.514*** GOV -3.264** -3.263* -3.285** -3.282* INF -6.17*** -5.89*** -15.115*** -15.812*** Note: *, ** and *** denote statistical significance at 1%, 5% and 1% levels respectively. For the Philippines, the first ARDL odel without any structural brea, yielded an F- value of only 3.42, which falls between the critical upper and lower bound [2.62, 3.79] at the 5 per cent level of significance. This suggests that the result obtained for the Philippines is inconclusive. The second ARDL odel with the crisis duy only, yielded an F-value of only 2.66, which falls below the critical lower bound of [2.86,
The Epirical Econoics Letters, 1(12): (Deceber 211) 1253 4.1]. This indicates that there is no steady state long-run equilibriu aong the variables in the second odel. The Philippines is always plagued by political instability, with frequent changes in governent. In 1986, Ferdinand Marcos fell fro power after reigning for ore than 2 years since 1965. He was reoved fro the office by the People Power vis-a-vis the EDSA (Epifanio de los Santos Avenue) Revolution, which is also nown as People Power Revolution or the Philippine Revolution of 1986. To tae into consideration, the political change that has taen place in the Philippines, the third ARDL odel for the Philippines is constructed by including another duy for political cliate. Once the political stability duy for 1986 to tae into consideration the political change that too place is included in the estiation, the overall result shows a drastic iproveent whereby, the calculated F-value (Wald test) increased to 7.3, and this value is higher than the critical upper bound of [2.86, 4.1]. This indicates that there is a steady state long-run equilibriu aong the variables in the case of the Philippines. Table 2: Results of Cointegration Test (F-Statistic Version of Bounds Test) (H : No Cointegration) Diagnostics Philippines Philippines (with Philippines (with crisis duy) political duy) Coputed F-Statistics 3.42 2.66 7.3* Note : The lower and upper critical values for the F-statistic version of the bounds test [lower critical bound, upper critical bound] at 5% significance level is [2.62, 3.79] and [ 2.86, 4.1]. * denotes the coputed test statistic (F-test) exceeds the upper critical bounds at 5% level of significance, thus rejecting the null of no cointegration. Case III of Pesaran et al. (21), that is unrestricted intercepts and no trends is followed. Table 3 shows the long-run estiated coefficients for the three different odels, naely without structural brea, with a crisis duy and with a political change duy. Once the crisis duy is introduced into the odel, the variable EG becoes significant at the 5 per cent level of significance and it has the expected negative sign. Variables FDI and INF are also found to be significant at 1 per cent level of significance. By including the political change duy into the equation, the overall odel showed iproveents. When the second duy naely the political change duy is introduced into the odel, VR, EG and GOV are found to be significant.
The Epirical Econoics Letters, 1(12): (Deceber 211) 1254 Table 3: Long-run Estiated Coefficients Variables Philippines (without structural brea) ARDL(,1,2,2,1) Coefficients (Dependent Variable: UR) Philippines (with crisis duy) ARDL(,,,3,) Philippines (with political change duy) ARDL(,2,2,2,1) VR +1.395 +.4637 +1.664*** EG +.66 -.2429** +.1951** GOV -1.8116* -1.9138 -.8765* FDI +3.4381** +3.973*** -.9493 INF +.5576 -.1787*** -.394 DUM1 - +1.844** - DUM2 - - +4.2656*** Note: *, ** and *** denotes statistical significance at 1%, 5 % and 1% levels respectively. One intriguing finding for the Philippines is that the positive coefficient of VR in third ARDL odel suggests the presence of structural uneployent in the Philippines. A positive sign of VR for the Philippines that contradicts with the theory of Beveridge Curve, iplies that the Philippines is facing a isatch proble. The VR was also found to be significant at 1 per cent level of significance in the third odel when the political change duy is taen into account. A isatch in the education and training output visa-vis the eployent requireent and utilization is cited as a ajor reason for increasing uneployent by Cruz (23). The author further added that this factor is ore doinant copared to expansions of the woring-age population and the econoy s inability to generate eployent. In the third odel, EG does not have the expected negative sign even though it is found to be significant. FDI plays an insignificant role in the econoy of the Philippines. Even though FDI has the expected negative sign in the odel, it is not significant, suggesting that FDI is not an iportant variable in the deterination of uneployent in the Philippines. In fact, the Philippines has one of the lowest investents in infrastructure in Southeast Asia, causing it to be a less attractive destination for FDI. The Philippines investent in infrastructure is only at 2-3 percent of its GDP (Balboa and Medalla, 26). This proble is further accentuated by the political instability. Soehow, the establishent of special econoic zones (SEZs) have substantially increased investents as well as exports in the Philippines (Maabenta, 22). Maabenta (22) also found that access to infrastructure, which lowers transportation costs is a ajor factor in the
The Epirical Econoics Letters, 1(12): (Deceber 211) 1255 deterination of FDI location in the Philippines. The geographical factor of the Philippines is also unattractive for FDI. Locations other than Luzon are less preferred by foreign investors, causing uneployent to reain high in other parts of the Philippines. The Palawan Island for exaple, has an uneployent of 19.6 percent in 1999 (Asian Developent Ban, 22). In the third odel, the variable INF is found to be insignificant but has the expected negative sign, suggesting that there exists a trade-off between uneployent and inflation. This finding is siilar to that of Kooros (26). The inclusion of a political change duy has two ajor iplications for the Philippines labour aret. Firstly, the isatch proble is enlightened in this study. According to Canlas (1997), the rising agricultural productivity in the Philippines releases worers fro this sector, and worers released fro the agricultural sector lac the sills required by the industry and services sectors, and, hence, suffer spells of uneployent. Li (1997) highlighted the inability of the Philippines econoy to absorb labour force participants that are ore silled and fro higher educational bacground based on the sall share of professional and technical worers and adinistrative, executive and anagerial worers, causing the to see eployent abroad. Galenson (1992) also highlighted that the principal cause of uneployent in the Philippine was the failure of the anufacturing sector to absorb the rapidly growing labour force. Second, political instability ay also increase uneployent as it reduces FDI as well as econoic growth. 6. Conclusion A positive relationship of uneployent with job vacancy suggests that an increase in job vacancy does not contribute towards a reduction in uneployent. We suspect that there could be a isatch proble in the Philippines labour aret. Thus, active labour aret policies are vital in the Philippines labour aret. It is also iportant for the Philippines to aintain a stable governent, as this, in turn, deterines FDI and econoic growth, that is crucial for eployent creation. References Abras, A. Burton, 1999, The effect of governent size on the uneployent rate, Public Choice, 99, 395-41 Asian Developent Ban, 22, Subregional Cooperation Strategy and Progra Status Report: Southeast Asia, http://www.adb.org/docuents/csps/serd/reg_in225_2.pdf (accessed on January 5, 27).
The Epirical Econoics Letters, 1(12): (Deceber 211) 1256 Assarsson, Bengt and Per Jansson, 1998, Uneployent persistence: the case of Sweden, Applied Econoic Letters, 5, 25-29. Balboa, D. Jenny and Erlinda, M. Medalla, 26, State of Trade and Investents in the Philippines. Discussion Paper Series No. 26-15, Philippines Institute of Developent Studies. Barro, J. Robert, 1991, Econoic growth in a cross-section of countries, Quarterly Journal of Econoics, 51, 47-443. Barsden, G., 1989, Estiation of long-run coefficients in error correction odels, Oxford Bulletin of Econoics and Statistics, 51, 345-35 Canlas, B.Dante, 1997, Uneployent and Monetary Policy in the Philippines, In Eployent, Huan Capital and Job Security: Recent Perspectives on the Philippine Labour Maret, ed. Eanuel F. Esbuerra and Kazuhisa, Ito. SEDP, No. 41, Institute of Developing Econoies. Christopoulus, K. Diitris and Efthyios, G. Tsionas, 22, Uneployent and governent size: Is there any credible causality? Applied Econoics Letters, 9, 797-8. Cruz, T.Teresa, 23, Uneployent in the Philippines: Is it really a structural proble? Institute for Labor Studies, Bureau of Local Eployent and Technical Education and Sills Developent Authority, Philippines. Downes, S. Andrew, 1998, An econoic analysis of uneployent in Trinidad and Tobago. Paper presented at Labour aret: Between Solidarity and Loyalty Seinar held at Montevideo, Uruguay fro May 19-2, 1998. Driffield, L. Nigel and Taylor, Karl, 2, FDI and the Labour Maret: A review of the evidence and policy iplications, Oxford Review of Econoic Policy, 16, 3: 9-13. Walter, Galenson, 1992, Labor and Econoic Growth in Five Asian Countries. New Yor: Praeger Kooros, K. Syrous, 26, In search of general uneployent odel, International Research Journal of Finance and Econoics, 4, 195-26. Layard, G. Richard and Stephen, J. Nicell, 1986, Uneployent in Britain Econoica 53, S121-S169. Li, Y. Joseph, 1997, Macro Labor Statistics and Sectoral Eployent, In Eployent, Huan Capital and Job Security: Recent Perspectives on the Philippine Labour Maret, ed. Eanuel F. Esbuerra and Kazuhisa, Ito. SEDP, No. 41, Institute of Developing Econoies.
The Epirical Econoics Letters, 1(12): (Deceber 211) 1257 Maabenta, P. Maria, 22, FDI Location and the Special Econoic Zones in the Philippines Review of Urban & Regional Developent Studies, 14(1), 59-77 Pesaran, M. Hashe, Yongcheol Shin, and Richard, J. Sith, 21, Bounds testing approaches to the analysis of long-run relationships, Journal of Applied Econoetrics, 16, 289-326. Edund, S. Phelps, 1994, Structural Slups: The Modern Equilibriu Theory of Uneployent, Interest and Assets. Cabridge:Harvard University Press. Ruane, P. Frances and Holger, Gorg., 1999, Irish FDI policy and investent fro the EU, In Innovation, Investent and the Diffusion of Technology in Europe: Geran Direct Investent and Econoic Growth in Postwar Europe, ed. Ray Barrell and Nigel Pain. Cabridge: Cabridge University Press. Scully, W. Gerald, 1989, The size of the state, econoic growth and the efficient utilization of national resources, Public Choice, 63, 149-164. Seyf, A., 2, Can ore FDI solve the proble of uneployent in the EU? A short note, Applied Econoic Letters, 7, 2: 125-128. Tang, Tuc Cheong, 22, Deand for M3 and expenditure coponents in Malaysia: assessent fro bounds testing approach, Applied Econoic Letters, 9, 72-75. Teo, Edward; Shandre, M. Thangavelu; and Elizabeth, Quah, 24, Singapore s Beveridge Curve: A coparative study of the uneployent and vacancy Relationship for selected East Asian countries, Econoic Survey of Singapore, Second Quarter. Trivedi, K. Pravin and Michael, G. Baer, 1985, Equilibriu uneployent in Australia: Concepts and easureent, Econoic Record, 61, 629-643. Valentine, To, 1993, The sources of uneployent: A siple econoetric analysis, Econoic Papers, 12, 1-2.