ANALELE ŞTIINłIFICE ALE UNIVERSITĂłII ALEXANDRU IOAN CUZA DIN IAŞI Tomul LVI ŞtnŃe Economce 2009 POLITICAL REGIME DURABILITY, DEVELOPMENT AND GOVERNANCE: THE ROMANIA S CASE Mha MUTASCU * Abstract The paper analyzes emprcally, n Romana s case, the relatonshps between poltcal regme durablty (dependent varable) and economc development & form of government (ndependent varables). The analyss s based on the constructon of a lnear Probt Model and the data set s coverng the perod 1926-2007. The man results show that, n Romana, the poltcal regme longevty s one n whch the democratc system s very strong, the economc development s hgh and the state s organzed n the form of monarchy. By excepton, a poltcal regme may have longevty even f the poltcal system s dctatoral, the economc development s nsgnfcant and the state organzaton (monarchy or republc) has mnmal nfluence. Key words: regme durablty, economc development, governance, probt analyss JEL classfcaton: H10, H11, O10, C35 1. Introducton Accordng to Marshall & Jaggers [2009, 16], the poltcal regme durablty represents the number of years snce the most recent regme change or the end of transton perod defned by the lack of stable poltcal nsttutons. Moreover, the researches of the causal relatonshps between poltcal regme durablty and economc development & form of government are not conclusve; some of them clam the connexons of the same sgn and other authors of the contrary sgn. In such a context, ths scentfc approach s ntended to analyze the relatonshp between poltcal regme durablty and ts determnant factors of economcal and poltcal nature. Based on the mentoned premse, all the theoretcal elements presented allow us to formulate a seres of theoretcal workng assumptons, whch consder two of the approachng coordnates of poltcal regme durablty: an economcal one and another one poltcal. * Mha MUTASCU (mha.mutascu@gmal.com), PhD, Assstant Professor, "The West Unversty of Tmsoara", Faculty of Economcs and Busness Admnstraton.
434 Mha MUTASCU 2. Theoretcal fundaments The feld lterature offer contradctory results about the sgn and the sense of the relatonshp between form of poltcal regme durablty, governance and development. Lpset [1959, 69-105] argues that the economc development ensures the democratzaton, more precsely development frst and democracy later. He consders that broadly based economc development s conducve to a democracy. Damond [1988, 3] beleves that f regme durablty vares wth economc performance, and dctatorshps are more able to channel resources to accumulaton and the creaton of wealth rather than consumpton, democratc regmes are more vulnerable to economc performance setbacks and poltcal turmol than ther authortaran counterparts. In the same note, Almond [1991, 467-474] revews key works of some scholars and explans, statstcally, the sgnfcant correlaton between economc development and democratc nsttutons (the accentuaton of economc development generates an ncrease of state democratzaton). Przeworsk [1991, 32] consders that To evoke complance and partcpaton, democracy must generate substantve outcomes: t must offer all the relevant poltcal forces real opportuntes to mprove ther materal welfare. Moreover, they suggest that economc development fosters democracy and promotes poltcal stablty. Resler and Kanet [1993, 5-22] consder that democraces buld ther legtmacy on nsttutonalzed procedures and consttutonal guarantees of poltcal rghts and freedoms, whle the prmary means through whch dctatorshps establsh ther legtmacy s good economc performance. In ths context, economc setbacks are more lkely to create nstablty n dctatorshps than n democraces. Preworsk and Lmong [1997, 155-183] show that the sustanable development can support the democracy survve, even f the countres are poorer. Pe [1999, 2] consders that the economc development wll transform socal structure and create a large enough mddle class as the socal bass of democracy. Przeworsk, Alvarez, Chebub and Lmong [2000, 78-88] develop ther prevous research and conclude: the economc development does not generate democraces, but democraces are much more lkely to survve n wealthy socetes. Chen [2007, 16-22], after he made a revew of the feld lterature, explans that the states wth hgh economc growth are strong democraces, havng the hghest level of development. Moreover, he formalzed the dea that the economc development s growng, as the democratzaton s ncreasng and vce-versa. Fnally, Robnson [2006, 1], analyzng the economc effects of development on democracy, concludes that the applcaton of technques adopted from best-practce econometrcs shows no evdence that economc development has a causal effect on democracy. More, nether does t support the dea that economc development nfluences the probablty of coups but not democratzatons. Therefore, the researches on the causal relatonshp s sgn between poltcal regme durablty and economc development & form of government are not conclusve; some of them clam the connexons of the same sgn and other authors of the contrary sgn. Ths scentfc approach s ntended to analyze the relatonshp between poltcal regme durablty and ts determnant factors of economcal and poltcal nature. Accordng to the mentoned premse, all the theoretcal elements presented allow us to formulate a seres of theoretcal workng assumptons, whch consder two of the approachng coordnates of poltcal regme durablty: one economcal and another poltcal one. The hypotheses are:
Poltcal Regme Durablty, Development and Governance: the Romana s Case 435 H1: The level of poltcal regme durablty s growng as the development s hgher. H2: The level of poltcal regme durablty s growng as the democracy s stronger or the autocracy s weaker. In summary, the meanngs of the hypothess work relatons are: The trend of poltcal regme durablty Table no.1 - The sense ( the sngs ) of the hypothess work relatons The determnant factors of poltcal regme durablty + 1. Development + + 2. Form of government - Democracy + + 3. Form of government - Autocracy - The trend of determnant factors of poltcal regme durablty The fundamental assumpton s that poltcal regme durablty represents a complex phenomenon determned by a couple of factors, such as: economc development and form of governance (democracy or autocracy). The lnkages are n the same sense for the economc development and democracy (not autocracy). 3. Methods and results Startng wth the theoretcal argues shown, the paper analyzes emprcally, n Romana s case, the relatonshps between poltcal regme durablty (dependent varable) and economc development & form of government (ndependent varables). The analyss s based on the constructon of a lnear Probt Model and the data set s coverng the perod 1926-2007. 1) The poltcal regme durablty (Regme Durablty Score - D) s quantfed by Marshall & Jaggers [2007] and represents the number of years snce the most recent regme change or the end of transton perod defned by the lack of stable poltcal nsttutons. 2) The economc development (per capta GDP - GC) s quantfed by Madson [2003] and Internatonal Monetary Fund [2009] and suggests the level of economc development as GDP per capta (1990 Internatonal Geary-Khams dollars). 3) The form of government (Combned Polty Score - PR) s taken from Marshall & Jaggers [2007] and the score scale ranges are from +10 (strongly democratc) to -10 (strongly autocratc). In ths study, the value of the ndependent varables D score becomes PD - Probablty of Poltcal Regme Durablty (the probablty that the poltcal regme durablty can exceed 4 years or not). PD values are 1 - when the poltcal regme durablty s greater than 4 years and 0 - f durablty s less than 4 years. Because the consdered ndependent factors (GC and PR) have dfferent scales of measurement, for a comparatve analyss, the levels of varables were normalzed: GC, PR GC, PR GC, PR = Max (1) Normalzed GC, PR GC, PR Max Mn GC [ 0,1] (2) Normalzed PR [ 1,0] (3) Normalzed
436 Mha MUTASCU In ths case, GC=0 ndcates a very hgh level of GDP per capta and 1 an extremely small one. PR= -1 s the level that corresponds to the strongly democratc states and 0 to the ones whch have a strongly autocratc regme. Based on the normalzed llustrated varables, the sense of changes exstng between probablty of poltcal regme durablty and ts determnant factors, accordng wth theoretcal assumptons made above, s as follows: Table no.2 - The expected sense ( the sngs ) of the relatons between PD - GC and PR accordng to workng hypothess The probablty of poltcal regme durablty The determnant factors of poltcal regme durablty The trend of determnant factors of poltcal regme durablty + GC - + PR - Moreover, I entered a dummy varable - T, whch reflects the type of the state (monarchy or democracy). If the state s a monarchy, the dummy s 1, and f the state s a republc, dummy s 0 (n Romana, n the consdered sample, the monarchc perod covers the nterval 1926-1947). Accordng to Dougherty [2007, 262], n probt estmaton, F(Z) - the standardzed cumulatve normal dstrbuton, gves the probablty of the event occurrng for any value of Z: p = F( Z ) (4) Maxmum lkelhood analyss s used to obtan estmates of the parameters. The margnal effect of x s δ p, whch s best computed as: δ x δp δp δz = x = f ( Z) xβ (5) δx δz δx where x represents the ndependent varables (GC, PR and T) and β the ndependent varables coeffcents. In ths case, the margnal effect of Z on the probablty, whch wll be denoted f(z), s gven by the dervatve of ths functon wth respect to Z: 1 2 1 Z 2 f ( Z ) = e (6) 2π As wth logt analyss, the margnal effect of any varable s not constant. It depends on the value of f(z), whch n turn depends on the values of each of the explanatory varables. To obtan a summary statstc for the margnal effect, the usual procedure s parallel to that used n logt analyss, basng of the mean values of the explanatory varables. In the consdered case, the Z s: Z = PD = α + β xgc + β xpr + β xt (7) 1 2 3 where α are the ntercept term and s the perod of tme (years 1926-2007). From 82 ncluded PD observatons, 32% s 0 (the poltcal regme durablty s less then 4 years) and 67% s 1 (the poltcal regme durablty s more then 4 years):
Poltcal Regme Durablty, Development and Governance: the Romana s Case 437 Dependent Varable: PD Method: ML - Bnary Probt (BHHH) Date: 06/05/09 Tme: 20:25 Sample: 1926 2007 Included observatons: 82 Frequences for dependent varable Table no.3 - The PD frequences n 1926-2007 perods Cumulatve Value Count Percent Count Percent 0 27 32.00 27 32.93 1 55 67.00 82 100.00 The econometrc tests of the Probt model are: Table no.4 - The econometrc tests of the Probt model PD - GC, PR and T Dependent Varable: PD Method: ML - Bnary Probt (BHHH) Date: 06/05/09 Tme: 20:25 Sample: 1926 2007 Included observatons: 82 Estmaton settngs: tol= 0.00010 Intal Values: C(1)=-1.70655, C(2)=-2.05023, C(3)=0.46344 Convergence acheved after 73 teratons QML (Huber/Whte) standard errors & covarance Varable Coeffcent Std. Error z-statstc Prob. GC -2.131737 0.623889-3.416853 0.0006 PR -2.315238 0.522454-4.431466 0.0000 T 0.497454 0.411948 1.207565 0.2272 Mean dependent var 0.670732 S.D. dependent var 0.472840 S.E. of regresson 0.436112 Akake nfo crteron 1.165287 Sum squared resd 15.02527 Schwarz crteron 1.253337 Log lkelhood -44.77676 Hannan-Qunn crter. 1.200638 Avg. log lkelhood -0.546058 Obs wth Dep=0 27 Total obs 82 Obs wth Dep=1 55
438 Mha MUTASCU The tests of model show the followng: the absolute values of the standard errors correspondng to the coeffcents of the functon are lower than the values of the coeffcents, wtch sustans the correct estmaton of these coeffcents (a concluson renforced by the low values of the probabltes); the value of the correlaton coeffcent - 67.07%, shows a sgnfcant statstcal correlaton between the dependent varable - PD and the ndependent varables - GC, PR and T; the Hannan-Qunn crteron (wth a resultng value under the crtcal pont of 2) shows that the resdual varables are low autocorrelated. In base of the model, the expectaton-predcton values are: Table no.5 - The expectaton-predcton values of PD n the base of the model Dependent Varable: PD Method: ML - Bnary Probt (BHHH) Date: 06/05/09 Tme: 20:25 Sample: 1926 2007 Included observatons: 82 Predcton Evaluaton (success cutoff C = 0.5) Estmated Equaton Constant Probablty Dep=0 Dep=1 Total Dep=0 Dep=1 Total P(Dep=1)<=C 11 0 11 0 0 0 P(Dep=1)>C 16 55 71 27 55 82 Total 27 55 82 27 55 82 Correct 11 55 66 0 55 55 % Correct 40.74 100.00 80.49 0.00 100.00 67.07 % Incorrect 59.26 0.00 19.51 100.00 0.00 32.93 Total Gan* 40.74 0.00 13.41 Percent Gan** 40.74 NA 40.74 The estmated model correctly predcts 80.49% of the observatons (40.74% of the Dep=0 and 100% of the Dep=1 observatons). Overall, the estmated equaton s 13.41% ponts better at predctng responses than the constant probablty model. In concluson, the model may be consdered representatve to descrbe, n Romana, the connecton between PD and GC, PR & T. 4. Conclusons The method for dentfyng the effect of the GC, PR and T on the probablty of PD conssts n calculatng the margnal effect at the mean value of the explanatory varables. The next table shows the margnal effects, calculated by multplyng f(z) by the estmates of the coeffcents of the probt regresson.
Poltcal Regme Durablty, Development and Governance: the Romana s Case 439 Table no.6 - The margnal effects of the Probt model PD - GC, PR and T Varable Mean β Mean β f(z) β x f(z) GC 0.739593-2.13174-1.57662 0.363592-0.77508 PR -0.8094-2.31524 1.873963 0.363592-0.8418 T 0.268293 0.497454 0.133463 0.363592 0.18087 Total Z (PD) 0.430809 Startng from the margnal effects measured on the probt model bult, we can dentfy the followng remarks: an one-pont ncrease n the GC, degreases wth 77.5% the probablty of poltcal regme durablty to be more than 4 years; an one-pont ncrease n the PR, degreases wth 84.1% the probablty of poltcal regme durablty to be more than 4 years; an one-pont ncrease n the T, ncreases wth 18.08% the probablty of poltcal regme durablty to be more than 4 years. We can observe that the results confrm the conclusons of Damond [1988], Przeworsk and Lmong [1991], Resler & Kanet [1993] and Chen [2007], but are n dsaccord wth the remarks of Robnson [2006]. A novelty s the exstence of a sgnfcant mpact of the type of state (monarchy or republc) on the poltcal regme durablty. For the analyzed perod, n Romana, an augmentaton n the level of economc development (GDP per capta) and an ncrease of the degree of democratzaton, on a monarchcal base, ncreases the probablty of poltcal regme durablty to be more than 4 years. Per a contraro, a dmnuton n the level of economc development (GDP per capta) and an ncrease of the degree of autocratzaton, on a republcan base, decreases the probablty of poltcal regme durablty to be more than 4 years. Between the three determnant factors (GC, PR and T), the most mportant s the degree of democratzaton, followed, wth a small dfference, by the level of economc development and state type (monarchy or republc). We can note that the poltcal stablty depend prmarly on the poltcal factors and than on the economcal determnants ones. In Romana, poltcal regme longevty s one n whch the democratc system s very strong, the economc development s hgh and the state s organzed n the form of monarchy. Accordng to the econometrc results, t appears that only a hgh level of socal welfare, combned wth a consttutonal monarchy and strong representatve democracy, allows premses to ensure a stable poltcal regme. In the same context, poltcal regme nstablty s treated as a level of reduced socal welfare, generated n the republcan perod, the full power of the state beng concentrated n the hands of a sngle person (autocracy). The forecast of the probablty of the poltcal regme durablty to be more then 4 years, n the 1926-2007 perod, n Romana, s llustrated n the follow graphc:
440 Mha MUTASCU 1.2 0.8 0.4 0.0-0.4-0.8 1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 The forecast of the probablty of poltcal regme durablty > 4 years (%) Graphc 1. The forecast of the probablty of poltcal regme durablty > 4 years (%) Based on the obtaned forecast probablty, we can observe the exstence of the strong negatve shocks (the probablty of the poltcal regme durablty to be more then 4 years s practcally mpossble) n the years: 1940 - the authortaran Kng Carol II has abdcated and was succeeded by the Natonal Legonary State, n whch power was taken by Ion Antonescu; 1944 - Antonescu was toppled and arrested by Kng Mchael I of Romana; 1947 - the communsts forced Kng Mchael I to abdcate and to leave the country, and proclamed Romana a republc and 1989 - popular uprsng aganst the Ceausescu regme and hs fall. A strong postve shock s observed n 1996, n whch a coalton of rght took the power. From ths year, the probablty of the poltcal regme stablty to be more then 4 years s rapdly growng and culmnates n the last year of analyss. Very nterestng s the hgh level evoluton of the probablty of the poltcal regme stablty to be more then 4 years n the communst perod, the fact explaned by the PR as a form of autocratcdctatoral government. The man results show that, n Romana, the poltcal regme longevty s one n whch the democratc system s very strong, the economc development s hgh and the state s organzed n the form of monarchy. By excepton, a poltcal regme may have longevty even f the poltcal system s dctatoral, the economc development s nsgnfcant and the state organzaton (monarchy or republc) has mnmal nfluence. References Almond, G.A., Captalsm and Democracy, Poltcal Scence and Poltcs, 24(3), 1991. Chen, L., Development Frst, Democracy Later? Or Democracy Frst, Development Later? The Controversy over Development and Democracy, The annual meetng of the Southern Poltcal Scence Assocaton, New Orleans, 2007.
Poltcal Regme Durablty, Development and Governance: the Romana s Case 441 Damond, L., Class, Ethncty and Democracy n Ngera: The Falure of the Frst Republc, London: Macmllan, 1988. Lpset, S., Some Socal Requstes of Democracy: Economc Development and Poltcal Legtmacy, Amercan Poltcal Scence Revew, 53, 1959. Maddson, A., Hstorcal Statstcs for the World Economy: 1-2003 AD, Horzontal fle, Copyrght Angus Maddson, 2009. Marshall, M., Jaggers, K., Poltcal Regme Characterstcs and Transtons 1800-2007, Polty IV Project, Center for Systemc Peace, 2009. Pe, M., Economc Insttutons, Democracy, and Development, Conference on Democracy, Market Economy, and Development, 1999. Przeworsk, A., Democracy and the market: poltcal and economc reforms n Eastern Europe and Latn Amerca, Cambrdge Unversty Press, 1991. Przeworsk, A., Lmong F., Modernzaton: Theores and Facts, World Poltcs 49(2), 1997. Przeworsk, A., Alvarez, M.E., Chebub, J.A., Lmong, F., Democracy and Development: Poltcal Insttutons and Well-beng n the World 1950-1990, Cambrdge Unversty Press, 2000. Resler, T. J., Kanet, R. E., Democratzaton: The Natonal-Subnatonal Lnkages, In Depth, 3, 1993. Robnson, J., Economc Development and Democracy, Annual Revew of Poltcal Scence, Vol. 9, 2006. *** Internatonal Monetary Fund, World Economc Outlook Database, Aprl 2009.