Managing Global Transitions International Research Journal issN 1854-6935 (online) EDITOR Egon Žižmond, University of Primorska, Slovenia ASSOCIATE EDITOR Robert D. Hisrich, Thunderbird School of Global Management, usa EDITORIAL BOARD Zoran AvramoviC, University of Novi Sad, Serbia Terrice Bassler Koga, Open Society Institute, Slovenia Cene Bavec, University of Primorska, Slovenia Jani Bek6, University of Maribor, Slovenia Heri BeziC, University ofRijeka, Croatia Vito Bobek, University of Maribor, Slovenia Guido Bortoluzzi, University of Trieste, Italy Branko BuCar, Pace University, usa Suzanne Catana, State University of New York, Plattsburgh, usa David L. Deeds, University of Texas at Dallas, USA Jeffrey Ford, The Ohio State University, usa Ajda Fošner, University of Primorska, Slovenia William C. Gartner, University of Minnesota, usa Tim Goddard, University ofPrince Edward Island, Canada Noel Gough, La Trobe University, Australia Andras Inotai, Hungarian Academy ofSciences, Hungary Muhammad Ishtiaq Ishaq, Global Institute Lahore, Pakistan Hun Joon Park, Yonsei University, South Korea Štefan Kajzer, University of Maribor, Slovenia Jaroslav Kalous, Charles University, Czech Republic Maja KoneCnik, University of Ljubljana, Slovenia Leonard H. Lynn, Case Western Reserve University, usa Monty Lynn, Abilene Christian University, usa Massimiliano Marzo, University ofBologna, Italy Luigi Menghini, University of Trieste, Italy Marjana MerkaC, Faculty of Commercial and Business Sciences, Slovenia Kevin O'Neill, State University of New York, Plattsburgh, usa David Oldroyd, Independent Educational Management Development Consultant, Poland Susan Printy, Michigan State University, usa Jim Ryan, University of Toronto, Canada Hazbo Skoko, Charles Sturt University, Australia David Starr-Glass, State University of New York, usa Ian StronaCh, Manchester Metropolitan University, uk Ciaran Sugrue, University of Cambridge, uk Zlatko ŠabiC, University of Ljubljana, Slovenia Mitja I. TavCar, University of Primorska, Slovenia Nada Trunk ŠirCa, University of Primorska, Slovenia Irena Vida, University ofLjubljana, Slovenia Manfred Weiss, Johan Wolfgang Goethe University, Germany Min-Bong Yoo, Sungkyunkwan University, South Korea managing and production editor Alen Ježovnik, University ofPrimorska, Slovenia editorial office University of Primorska FaCulty of Management Koper Cankarjeva 5, si-6104 Koper, Slovenia Phone: ++386 (0) 5 610 2021 E-mail: mgt@fm-kp.si www.mgt.fm-kp.si indexing and abstracting Managing Global Transitions is indexed/ abstraCted in the International Bibliography of the SoCial SCienCes, EConLit, doaj, EConPapers, Index CoperniCus, Cabell's, ebsco, and ProQuest. Managing Global Transitions International Research Journal VOLUME 12 ■ NUMBER 4 ■ WINTER 2014 ■ ISSN 1854-6935 Table of Contents 303 Nonlinear Co-Integration Between Unemployment and Economic Growth in South Africa Andrew Phiri 325 Behind the Exporters' Success: Analysis of Successful Hungarian Exporter Companies From a Strategic Perspective Annamaria Kazai Onodi and Krisztina Pecze 347 Differentiation in Police Services in City Neighborhoods Uriel Spiegel, Tchai Tavor, Simon Hakim, and Erwin A. Blackstone 371 Stock Market Consequences of the Suspension of the Central Bank of Nigeria's Governor Ibrahim Mohammed and Chioma Nwafor 395 Investment and Profits: Causality Analysis in Selected eu Countries Igor Stubelj 415 Abstracts in Slovene Nonlinear Co-Integration Between Unemployment and Economic Growth in South Africa Andrew Phiri North-West University, South Africa phiricandrew@yahoo.com In this paper, a momentum threshold autoregressive (mtar) model is used to evaluate nonlinear equilibrium reversion between unemployment and economic growth for South African data between the periods 2000-2013. To attain this objective we estimate the first-difference and the gap model variations of Okun's specification. For the latter model variation, we employ three de-trending methods to obtain the relevant 'gap' data; namely, the Hodrick-Prescott (hp) filter, the Baxter-King (вк) filter and the Butterworth (bw) digital filter. A common finding from our empirical analysis is that Okun's law holds concretely for South African data regardless of the model specification or the de-trending technique that is used. Moreover, our analysis proves that unemployment granger causes economic growth in the long-run, a result which may account for the jobless-growth phenomenon experienced by South Africa over the last decade or so. Key Words: unemployment, economic growth, Okun's law, South Africa, mtar model, nonlinear unit root tests, nonlinear co-integration, nonlinear Granger tests, Hodrick-Prescott filter, Baxter-King filter, Butterworth high-pass filter j el Classification: C22, C51, E23, E24 Introduction High economic growth in conjunction with low unemployment under a low inflation environment can be deemed as the ultimate objective of macroeconomic policy in South Africa. Over the last decade or so, two prominent macroeconomic policy frameworks have embodied these objectives, those being, monetary policy's 'inflation-targeting' regime and fiscal policy's Accelerated and Shared Growth Initiative of South Africa (asgisa). Implemented in February 2002 and still in use to date, the inflation-target policy rule specifies that the South African Reserve Bank (sarb) should contain inflation at levels of between 3 and 6 percent, whereas the asgisa initiative seeks to halve unemployment and attain a Managing Global Transitions 12 (4): 303-324 6% economic growth rate by the year 2014. The assumed compatibility of the aforementioned policy objectives is inevitable demonstrated as monetary policy in South Africa is designated towards manipulating nominal variables like interest rates and inflation as a means of influencing real variables such as output growth and employment. Ultimately, the success of disinflation policy is reflected in its effect on unemployment and output growth. However, up-to-date South Africa has not only to managed to achieve arguably the highest economic growth rates in Africa since 1994, but the economy simultaneously boasts one of the highest youth unemployment rates in the world. So even though the South African Reserve Bank (sarb) can be credited for containing inflation within its set target which has been accompanied with steadily improved economic growth, such acquired growth has been characterized by what is popularly referred to as a 'jobless growth' syndrome (Hodge 2009). A mystery is warranted since the 'jobless growth' phenomenon contradicts the epic rise of unemployment caused by the sharp decline of real output experienced worldwide during the great depression. Therefore, a classical challenge for academics and policymakers alike is to provide an adequate account of unemployment-growth correlations in the South African economy. The question regarding the linkage between economic growth and unemployment gained prominence after Okun (1962) depicted the extent to which the unemployment rate is negatively correlated with output growth. By analyzing data over the period of 1947 to 1960, Okun (1962) documented that unemployment in the United States tends to fall by a one percentage point for every 3 percentage point rise in output growth. Thereafter, the United States was dubbed as having an estimated 'Okun coefficient' of 3 and a plethora of subsequent authors sought to estimate Okun's coefficient by either adopting a single-country approach (see Caraiani 2010; Ahmed, Khali, and Saeed 2011), panel-data approach (see Dixon and Shepard 2002,1997; Lal et al. 2010) or a multi-regional approach (see Freeman 2000; Adanu 2005; Villaverde and Maza 2009). The appeal of Okun's relationship is attributed to its simplicity and its extensive empirical support qualifies it to belong at the core of modern macroeconomics (Jardin and Gaetan 2011). As noted by Silvapulle, Moosa, and Silvapulle (2004), estimating the Okun coefficient has important implications for the business cycle since it relates the level of activity in the labour market to the level of activity in the product market. Whilst Okun's law implies that more labour is typically required for increased productivity levels, Okun's coefficient serves as an indication of the cost of unemployment in terms of output growth (Noor, Nor, and Ghani 2007). And in consolidation with the Phillips curve; Okun's relationship assists macroe-conomic policy in determining the optimal or desirable growth rate as a prescription for reducing unemployment (Moosa 1997). Overall, Okun's law is recommended as 'a rule of thumb' which provides policymakers with an understanding of how different markets adjust, and thus allowing for correct policies to be selected when facing shocks (Pereira, Bento in Silva 2009). In reality, Okun's law is more of a statistical relationship rather than a structural feature of the macroeconomy (Knotek 2007). The development of a pure theoretical foundation for Okun's relationship has been largely neglected in the academic literature, such that empirically, no functional form has been dominantly preferred to any other on the basis of theory (Weber and West 1996). As a consequence, the empirical examination of Okun's law is typically subject to revisions with the co-movement between output growth and unemployment frequently being analyzed under different settings. So while there is no contention on the importance of Okun's law, debates have evolved on the econometric techniques used to establish this relationship; how the cyclical components are extracted; and whether a dynamic or static specification is adopted (Turturean 2007). Recently, the possibility of asymmetric behaviour between economic growth and the unemployment rate has added a new dimension in the development of the academic literature. Take for instance Jardin and Gaetan (2011) who consider asymmetries in Okun's relationship as being important because asymmetric behaviour can adequately account for the varying effectiveness of structural and stabilization policies. Other commentators, such as Geldenhuys and Marnikov (2007), consider the impact of asymmetric behaviour on policy forecasting practices. In particular, these authors argue that if Okun's relationship is indeed found to be asymmetric, forecasts based on linear estimates of Okun's coefficient can lead to biased error terms. And yet another cluster of authors can also be identified, who advocate on the necessity of incorporating asymmetries in Okun's relationship as a means of reinforcing asymmetric behaviour in the Phillips curve. The rationale behind this line of thought is that if Okun's coefficient changes between regimes, then the sacrifice ratios should also change between regimes. In other words, different degrees of gradualism in the disinflation process may imply different im- pacts on unemployment for the same reduction in inflation (Beccarini and Gros 2008). Our study contributes to the literature by addressing the economic significance of asymmetric behaviour in Okun's relationship for South African data. To this end, our study makes use of the momentum threshold (mtar) autoregressive framework of Enders and Granger (1998). The logic behind the choice of our adopted approach can be described as follows. Engle and Granger (1987) argue that evidence of unit roots between a pair of time series variables necessitates the use of co-integration analysis prior to the estimation of any regression formed by the variables. According to the authors, the presence of co-integration would then imply that the variables follow a common long-run trend and the ols estimation of the time series will not yield spurious results. This is an important implication for our case study since previous empirical works have cautioned of unit root I(1) behaviour in output growth and unemployment variables for South African data (see Hodge 2006; Burger and Marnikov 2006; Gupta and Uliwingiye 2010). And yet it should also be noted that these conclusions are based on studies which assume a linear data generating process (dgp) among the series. Of recent, it has become widely accepted that standard unit root tests, suffer from low power when a linear approximation of an otherwise nonlinear time series is used to evaluate the integration properties of a time series (Enders and Granger 1998). A similar contention has risen for co-integration analysis, in which researchers like Enders and Dibooglu (2001) prove that the implicit assumption of symmetric adjustment is problematic if the adjustment towards long-run equilibrium is not linear. In particular, the authors argue that the presence of nonlinearities between a pair of time series signifies a high probability of nonlinear adjustment processes towards the long-run equilibrium for the data. With this in mind, our paper probes into the possibility of asymmetric behaviour between the unemployment rate and output growth using the mtar model. We choose this model because it represents a simple yet flexible framework that can simultaneously facilitate for (1) nonlinear unit root tests, (2) nonlinear co-integration analysis; and (3) nonlinear causality analysis. Therefore, against this backdrop, we present the remainder of the paper as follows. The following section of the paper presents the empirical framework of the study whereas section four presents the empirical results of the study. The paper is concluded in section five by providing policy recommendations and suggesting avenues for future research. Empirical Framework Our paper uses two classes of Okun's law specifications; namely, the first differences model and the gap model. To ensure that we obtain a balanced, robust vew on the estimation results, we specify the Okun's specifications on both the direct and the reverse regressions of unemployment on output growth. For instance, in specifying the 'first differences' version of Okun's law, the link between the unemployment rate (ur) and economic growth (gdp) is represented as: ' Agdpt^ / \ ßi 0 / \ Aurt + / \ 61 к Aurt j , 0 в j ч Agdpt / Л12, where A is the first difference operator such that Agdpt = gdpt - gdpt-i and Aurt = urt - urt-1. On the other hand, the 'gap model' measures these variables in terms of their deviations from long-run trends and is specified as: ' gdpcts / \ ßi 0 ( C \ urt + / \ ,urt J , 0 в2 J ,gdpt , where urct = urt - ur*t and gdpct = gdpt - gdp* are representative of the cyclical components of the unemployment rate and real output, respectively; with gdp* denoting a measure of potential output gap and ur* the unemployment gap variable. Having specified our baseline theoretical models, we can proceed to introduce co-integration analysis amongst the variables. We, therefore, take heed of Enders and Granger (1998) and model asymmetric adjustment between the unemployment and real output growth variables by allowing the residual deviations (i. e. £ti) from the long-run equilibrium of regressions (1) and (2) to behave as a tar process. Formally, these residuals are modelled as follows: p = ItPi£-i + (1 - It)P2&-1 + J]ßiAgt-i + st. (3) i=i In our paper, we identify four types of co-integration relations which govern the asymmetric dynamics within Okun's law, namely; tar with a zero threshold; consistent tar with a nonzero threshold; mtar with a zero threshold; and consistent mtar with a nonzero threshold. In the tar model with a zero threshold, the indicator function, It, is set according to: Under the tar model with a nonzero threshold, we set It, as: It i, if \&-i > t o, if gt-i < t (5) where t is the value of the threshold variable. Enders and Granger (1998) suggest the use of a grid search procedure, as demonstrated in Hansen (1997), to derive a consistent estimate of the threshold i. e. the threshold estimate yielding the lowest rss is considered the true threshold estimate. The tar models are designed to capture potential asymmetric deep movements in the residuals if, for example, positive deviations are more prolonged than negative deviations (Enders and Dibooglu 2001). Enders and Granger (1998) and Caner and Hansen (2001) suggest that by permitting the Heaviside indicator function, It, to rely on the first differences of the residuals, Agt-1, a mtar version of equation (11) can be developed. The implication of the mtar model is that correction mechanism dynamic since by using Agt-1, it is possible to access if the momentum of the series is larger in a given direction relative to the direction in the alternative direction. In other words, the mtar model can effectively capture large and smooth changes in a series whereas the tar model shows the 'depth' of the swings in equilibrium relationship. In modelling mtar threshold co-integration with a zero threshold, the indicator function Mt, is set as: While in the mtar model with a nonzero threshold, Mt, is set as: For both tar and mtar specifications, Enders and Silkos (1998) demonstrate that a sufficient condition for stationary of gt-1 is that p1, p2 < 0. If gt-1 is found to be stationary, the least squares estimates of p1 and p2 have an asymptotic multivariate normal distribution for any given value of a consistently estimated threshold. Moreover, the null hypothesis of no (6) (7) co-integration (i. e. н01: p1 = p2 = 0) can be formally tested using a standard F-statistic for both tar and mtar models. If the null hypothesis of no co-integration is rejected, it is possible to test for the null hypothesis of symmetric adjustment (i.e. h02: p1 = p2) against the alternative of asymmetric adjustment (i. e. н12: p1 * p2) using a similar F-test. The empirical F-distribution for the null hypothesis; p1 = p2 = 0is tabulated in Enders and Dibooglu (2001) whereas Enders and Siklos (2001) report critical values for testing the null hypothesis of p1 * p2. If both null hypotheses of no co-integration and no asymmetric co-integration can be simultaneously rejected, the granger representation theorem is satisfied and thus an associated error correction model can be estimated for the pair of time series variables. Thus in validating the presence of threshold co-integration, the error correction model can be modified to take into account asymmetries as in Blake and Fomby (1997). In our study we augment each of our threshold co-integration regressions with thresholds error correction specifications. In particular, the tar-tec model can be expressed as: Agdpt Aurt = AiJt£t-i + Л2 (1 - It)£t-i p p + 2 aiiAgdpt-i + ßiiAurt-1. (8) 1=1 Whereas the mtar-tec model is specified as: Agdpt Aurt = X21Mt 6-1 + ^22(1 - Mt )£-1 pp + ^ a2iAgdpt-1 + ^ ß2iAurt-1, (9) i=1 where the indicator functions for the tar and mtar model specifications are represented by It and Mt respectively. Through the above described systems of error correction models, two types of joint hypotheses can be tested. Firstly, the presence of asymmetries between the variables could initially be examined by examining the signs on the coefficients of the error correction terms. This involves testing the null hypothesis of H03: Ah6t-1 = Äi2^t-1 against the alternative н^: Л16-1 * Л-2&-1. The second type of hypothesis tested is that of granger causality effects which relatively examines whether all Agdpt-k and Aurt-k are statistically different from zero. Granger tests are used to examine whether the lagged values of one variable do not improve on the explanation or 'granger-cause' another variable. In particular, the null hypothesis that urt does not lead to gdpt can be denoted as: h04: a = 0, i = 1,..., k; whereas the null hypothesis that gdpt does not lead to urt is: h05: ßi = 0, i = 1,..., k. All aforementioned hypotheses are based on a standard F-test. Furthermore, three types of joint hypotheses can be formed from the tec model. Firstly, granger causality tests can be implemented by testing whether all Agdpt-k and Aurt-k are statistically different from zero based on a standard F-test and if the Л coefficients of the error correction are also significant. Empirical Analysis empirical data The data used in the empirical analysis consists of the annual percentage change in the real gross domestic product which is gathered from the South African Reserve Bank (sarb) online database whereas the unemployment rate for all persons aged above 15 years of age is collected from various issues of the quarterly labour force surveys (qlfs) as complied by Statistics South Africa (statssa). Our empirical analysis uses quarterly adjusted data obtained for the periods extending from 2000 to 2014. The choice of our sample period and periodicity reflects the limitations in the availability of the time-series data on unemployment and economic growth for South Africa. Although it would be desirable to employ a longer span of data, the available data provides the advantage of avoiding the issue of potential structural breaks related to South Africa's political and structural reforms such as those experienced in 1994. Moreover, we take note that while our data is relatively short, it is, however, up-to-date and further eliminates the problem of data unreliability associated with the South African unemployment series before 2000. Further given that gross domestic product is available on a quarterly basis and the unemployment rate is limited to half-yearly data, we use cubic spline interpolation to convert the half-yearly unemployment data into quarterly data over the same time period. We favour the use of cubic spline interpolation over other time series data conversion techniques due to its computational accuracy and stability of computation. Moreover, cubic spline interpolations satisfy the further condition at the end point. As a part of our data construction, we introduce the de-trending meth- ods used to extract the 'potential output' and 'unemployment gap' variables necessary to estimate the gap version of Okun's specification. The construction of these 'gap variables' is necessary since there exists no observable data on the trend components of the unemployment and output growth variables. Also taking into consideration that a majority of these de-trending techniques are not without scepticism, it is standard practice to apply a variety/different de-trending techniques to ensure robustness in the regressions analysis. Therefore in following along this course of reasoning, our study considers three alternative de-trending techniques, namely the Hodrick-Prescott (hp) filter; the Baxter-King (bk) filter and the Butterworth (bw) digital filter as respectively introduced by Hodrick and Prescott (1997), Baxter and King (1999); and Pollock (2000). The purpose of using these three de-trending techniques is to enable a robust analysis concerning the sensitivity of the estimated Okun's coefficient to the different choices of our gap variable estimates. unit root tests In testing for unit roots, we begin on the simple premise of subjecting a univariate time series, yt, to the following generalized autoregression: Yt = 0) + ut, (11) Ayt = St (st-i < t) + St (st-i > t) + ut, (12) Whereas the mtar version of the unit root test regression with a zero threshold and a consistent threshold estimate threshold are, respectively, specified as: Ayt = St(ASt-i < 0) + St(ASt-i > 0) + ut, (13) table 1 Nonlinear Unit Root Tests Variable Model Lag Asymmetry test Unit root test Decision (i. e. p, = p2) (i. e. p1=p1 = 0) gdp TAR 2 0.94 12.63*** Linear I(o) (3.32)* (16.46)*** Nonlinear I(o) c-TAR 2 З.94* 15.59*** Nonlinear I(o) (7.87)* (21.28)*** Nonlinear I(o) MTAR 2 0.95 12.13*** Linear I(o) (9.46)** (22.89)*** Nonlinear I(o) c-MTAR 2 4.90* 16.03*** Nonlinear I(o) (6.67)* (19.96)*** Nonlinear I(o) ur TAR o 2.45 2.86* Linear I(o) (4.96)* (7.22)** Nonlinear I(o) c-TAR o 2.37 2.81* Linear I(o) (5.21)* (7.40)** Nonlinear I(o) MTAR o 2.59 2.94* Linear I(o) (3.44)* (6.17)** Nonlinear I(o) c-MTAR o 2.70 3.00* Linear I(o) (3.37)* (6.12)" Nonlinear I(o) notes Significance level codes: ***, ** and * denote the 1%, 5% and 10% significance levels respectively. Tests statistics for the first differences of the variables, i. e. Agdpt and Aurt are given in parenthesis. Ay t = et (Aet-1 < т) + et (Aet- > т) + v t. (14) Thereafter, two hypotheses can be formed from regressions (ii)-(i4). The first hypothesis tests for asymmetries within the time series. To this end, we test the null hypothesis of no asymmetric effects as н00: p1 = p2 against the alternative hypothesis of an asymmetric data generating process (i.e. н01: p1 Ф p2). Subsequent to testing for asymmetric effects, we then proceed to test for unit root behaviour within the time series. Pragmatically, the null hypothesis of a unit root is tested as н10: p1 = p2 = 0 against the alternative hypothesis of an otherwise stationary asymmetric process (i.e. н10:p1 Ф p2 Ф 0). The aforementioned tests of asymmetry and unit root behaviour are performed on time series variables of economic growth and the unemployment rate. The lag length of the threshold models which facilitate these tests are determined by the aic information criterion. As is evident from table 1, the empirical test results obtained for the time series in their levels are quite mixed. For instance, in scanning through the model tests conducted on the unemployment variable, we find that we cannot reject the null hypothesis of a symmetric process and yet we are able to reject the null hypothesis of a unit root process for same time series. Thus for the unemployment variable in its levels, we conclude a linear, stationary data generating process for the series. However, for the output growth variable in its levels, we conversely find that the c-tar and c-MTAR versions of the employed tests simultaneously reject both null hypotheses of symmetry and unit root behaviour. This particular result implies a nonlinear, nonstationary data generating process for the output growth variable in its levels. And yet, in turning to the empirical results obtained for the time series in their first differences, our analysis reveals a common finding of a nonlinear yet stationary process for all variables under all model specifications. All in all, we can conclude that all utilized time series appear to be both nonlinear yet stationary processes in their first differences. Therefore, the results obtained from our preliminary unit root analysis paves the way for the threshold co-integration analysis which we conduct next. CO-INTEGRATION ANALYSIS Having investigated the integration properties of the unemployment and economic growth variables, we proceed to investigate threshold co-integration and error correction effects amongst the times series. However, prior to estimating any threshold models, we must first test a number of hypotheses to select which models best capture asymmetric behaviour in Okun's specification. To this end, we employ three threshold tests which have been previously discussed previously discussed. To recall, (1) we test for co-integration effects; (2) we test for threshold co-integration effects and (3) we test for threshold error correction effects. The results of these tests are reported in table 2. In referring to these results, we find that at least one type of threshold model manages to reject all three hypotheses at least a 10 percent significance level for all variations of Okun's law. This is quite an encouraging result since it implies that the data displays at least one form of nonlinearity for each version of Okun's specification. Another interesting result is that the m tar specification is most suitable for modelling nonlinear behaviour between unemployment and economic growth for South African data. The only exception holds for the CF filter estimates which favour a tar model specification. Furthermore, all estimated versions of Okun's law unveil significant asymmetric co-integration behaviour only when output growth is placed as the dependent variable in the regression. In summing up the test results reported in table 2, we can draw two broad conclusions thus far. Firstly, our analysis infers significant asymmetric behaviour between unemployment and economic growth for South African data. In this respect, our results adhere with those obtained in Geldenhuys and Marnikov (2007). However, in slightly differing from Geldenhuys and Marnikov (2007), we find smooth nonlinear adjustment behaviour in the data as opposed to an abrupt one. This result is expected since the otherwise abrupt nonlinearity is most suited for data containing structural break periods. Seeing that our data does not cover such periods, it therefore becomes reasonable that we detect smooth nonlinear behaviour among the data. Our second conclusion is that we establish economic growth as being the driving variable in the asymmetric relationship detected between the time series. This is worth observing since it serves as a guideline on how to estimate each of the selected threshold regressions. In our instance, we specify the mtar models under the assumption that economic growth is regressed on the unemployment rate. This is of course with the exception of the cf filter regression in which we model tar nonlinearity and yet retain economic growth as the dependent variable in the regression. Our estimation results of the first difference model specifications are reported below in table 3 whereas the results obtained for the gap model versions are reported in table 4. Starting with the results reported in table 3 for first differences model, we take note of a long-run coefficient estimate of -0.09. Technically speaking, the magnitude of this coefficient estimate as obtained under both first difference models implies that a 1 percent decrease in the unemployment rate is associated with a -0.09 percent increase in productivity output. This result is seemingly plausible as it does not violate traditional theory of a negative unemployment-growth co-relationship as initially postulated by Okun (1962). Furthermore, the magnitude of this relationship is consistent with some of the Okun coefficient estimates obtained in previous studies. Among these previous studies are the works of Adanu (2005) who obtain a similar estimate of -0.09 for Alberta province in Canada; Villaverde and Maza (2009) who find a -0.08 estimate for a regional group of Spanish data and also Geldenhuys and Marnikov (2007) who obtain an estimate of -0.11 for South African data. In moving on to examining the regime switching behaviour among the co-integration error terms, we firstly note that all threshold estimates table 2 Threshold Cointergation and Error Correction Tests Model (1) (2) tar-tec mtar-tec H(1) Ho H(2) Ho H(3) Ho H(1) Ho H(2) Ho H(3) Ho First differences Agdpt Aurt 25.36 (0.00)*** 4.10 (0.05)* 0.47 (0.50) 32.71 (0.00)*** 9.16 (0.01)** 2.47 (0.13) Aurt Agdpt 41.82 (0.00)*** 0.68 (0.42) 0.01 (0.91) 50.82 (0.00)*** 1.66 (0.21) 0.01 (0.95) HP filter gdpt ur A - (9') We assume that government ignores the differences in preferences, and treat identically all constituents, like in case 1. We assume also that all 3 0 local governments in the metropolitan area have the same cost function as above: T = CG to determine the same level of (6'). ^ N • A - C t C G =- = A-- N N and charge all customers the equal cost sharing burden as follows: C P, = Jf. (.0 For any customer we can define his consumer surplus as: „ t N / C \ N. C P,-=A + --i-A-- =--!+-. n 2 \ N) 2 N Therefore the total consumers' surplus in case 2 is: ^.z^-z^-N). - !=1 !=1 X ' Or, rrc (a^N\(A N \ N . N(l + N)(2N + l) TCS2 = (A + -|(A -- -i|. - +--- (12') 2 2 2 12 -i. * + £ 2 \2 N N C 2C — + — \\N +--4+1 2 N \ N 41 For simplicity of exposition, let us define three values of W, X and Y as follows: v Nil + N)W + i) X = -, (14) 12 rsi. * + £ 2 \ 2 N N C 2C — + — ]\N +--4+1 2 N \ N 1 (15) Thus, TCS2 is: TCS2 = W + X-Y-cU-jj\. (16) Comparing TCS1 to TCS2 yields (17): TCS, = (Ж ~ C)2 I W + X - 7 - CIA + -) = TCS2. (17) 2N 2 From (7') and (17) we can see that TCS1 - C (A + f) > W. The other two values at (17) are X and Y that have different signs. We can see very easily that for large population N the value ^ approaching zero, i. e. ^ —» o. Therefore it is most likely that X > Y, this is because: N3 N --N2 + — reservation price > OA + BN will benefit from the private security supply which will yield additional consumers surplus. The demand of the 'borderline' customer will not yet purchase private security. At this stage we want to find the extra net social welfare of all customers who buy the private security, PS, in addition to the supply of the public good. For this purpose we introduce first the 'leftover' demand for the private security and extra units to the public good supplied. For this purpose we introduce first to the leftover demand of the highest demand customer that is as follows: P= N C A + — — A-- 2 N _ N C ~ 7 + N' At equilibrium we can measure the quantity demanded for private security of this customer as follows. His demand is: 4M)- Since P = C we find the highest PS of this customer as For each other customer, I, the demand for PS is Pi = N C A +--1 — A-- 2 N 2N (21) The term in the first left bracket is the highest reservation price ofindi-vidual i for security and the second term in the bracket is G0 the optimal supplied quantity of the pure public good. For each customer, i, the 'leftover' demand for the private security, PS, is obtained as follows: N C Pi = AH— — A - — -i 2 N - i - PS - C. From (22) we find that PS, the demand for extra private security for each customer i, is: PS=^ + --i-C. (23) N 2 As i is higher, PS purchase of customer i is smaller, and PS is approaching zero when PS = o = ^- + --i-C, (24) N2 or C N — + — N 2 i = - + - - C. (24') The extra consumer surplus of each customer I, for consumer i (i = 5,1,2,...,(f + £-C)) is: C2 22 1 = 0 Therefore, we get (11') from (12') and (25) the extra/additional welfare resulting purchase of private security as supplement to the optimal public security, G0 as follows: AW= J] Senl = - ----l-, (25') or, AW = C N 2 C N — +--C -2 — +--C + i N2 ) \N 2 / C C \2 / C N ^ - — +--C -2 — +--C + i \ N N \N 2 (25'') 2 Therefore, the extra welfare obtained by private security supplement is + 4 The total extra private security purchase by all customers in addition to the mutual consumption of the pure public good, G0, is measured as follows: The PS of the highest demand for private security is PS = CN — +--C N 2 The next one is C r — +--C - 1 N 2 The last customer who prefers only the pure public security with any extra private supplement is customer C N ^ — +--C N 2 Thus, the total private security supplement units TPS, are TPS = PS CN — +--C N 2 CN — +--C + 1 N 2 (27) or TPS = Ш + f-cf (28') In the next stage we want to investigate the mixture between total private security purchased unit and the total public security unit. The ratio 'mixture' between the two kinds of security units are: Mix = TPS G0 (28) In the next section we examine how in case 3 changes in the independent variables A, C, and N affect the dependent variables, G0, TPS, Mix, and AW. 2 2 Comparative Static Analysis c- ^ л C u dG Since G = A--then — > o, N dA dG i . dG C — =--< o, and -— = — >o. dC N dN N2 Tp, № + f-C)2 ^ dTPS Since TPS =-then, —-— = o, 2 dA dTPS *(S + g-c)a-(M dC 2 Assuming N » A, and N » C and always A > C, we get that jj —> o. Therefore, dTPS t N ч = --C (-i)co, dC 2 dTPS + 7 + dTPS /N \ 1 - => —— =--C • - > o. dN 2 dN 2 2 * Ж < 0 More public security relatively to private supplement when reservation price is higher. * < 0 Lower production cost per unit of supplied security increases the mixed supply between private security and public security, i. e., less public security relative to private security (see appendix i). * > ° 'n most cases a larger community necessitates a larger mixture of private security in comparison to public security (see appendix 2). Taking the derivatives of equation (14) on changes in the Mix values with respect to A, C and N yields the following dMix (-2) Ш + f - cf [£ + f - C] - <0. dA Assuming N » A, N » C and always A > C,we obtain dMix A - C N —^7- ~--:---<0 dC 4A2 2 2 table 1 Independent variable Dependent variable change dMix dG dTPS d AW dA - + 0 0 dC - - - - dN + + + + From (26) and (27') we find that AW is equal to Therefore, the signs of the values and ^f are similar to the signs of the par- allel values = =<, and ^ > o. All the results above are summarized in table 1. Based on table 1 we determine several additional results regarding the effects of A, C and N on the optimal values of public and private security expenditures and the welfare effects. Higher value of A that indicates a larger 'necessity' for security leads to higher spending on public security. However, it does not affect the private security supplement expenditures that are spent by each private individual since it is cheaper to finance security publicly Moreover, higher values of A reflect higher social welfare from public security. However, it does not change the values of TPS, Total Private Security. To summarize, more requirements for security increase permanently the value added of social welfare, leading to a Pareto improvement. The effects of increase in C, the production cost of security, on the decision variables are straightforward. It reduces the attitude to spend money on security of any kind, public or private. However, higher C of individually paid private security is more significant than collectively paid public security. Thus, the Mix decreases too. The most important results in our comparative static analysis relates to the population size, N, and has two contradicting effects. On the one hand, the increase in N increases the advantage of cooperation among consumers of sharing the burden of public spending. Moreover, the larger the population, N, the greater is the advantage of purchasing of more public security On the other hand, the increase in N leads to groups with different demands for security. Therefore, in the specific rectangular distribution of demand and taste, public security increases by a lower percentage than private security. This leads to a higher mixture, Mix, and to a consistent increase in the importance of supplementing public police. We may predict based on our results that in larger and more diversified communities the supplement of private security is more significant, while in more homogeneous or small communities the population may rely on collective public security than in other large and non homogeneous communities. The requirement that N » C is crucial since it emphasizes the possibility of sharing the burden of public security before paying privately and individually for private security. A high level of C, may discourage many individuals from buying private security and from consistent increase in welfare. Supplementing Public Police The budgeting process of local government addresses the preferences of the 'median voter' which is determined through the political system. Specific preferences could be addressed by group of residents when their number reaches the economic threshold size. When the local political pressure to address the specific growth of services by the locality as a 'public good' is perceived difficult or fails then the group resorts to group effort like (1) private police or (2) volunteer effort like vigilante groups or specifically neighborhood watch. When such group action is difficult or involves high transaction costs by the individuals who wish to supplement public police then individual efforts are employed. Such individual activities include (3) self-protection, (4) property insurance, and (5) protection design. Self-protection includes deterrence, prevention, and detection measures. Individual preferences may motivate others to resort to acquire insurance policies with lower deductibles and greater coverage. The last security measure that is available typically when properties are built is environmental where access through windows and doors is made difficult, and access is controlled. Private Police Private police are estimated to be at least three times the combined federal, state, and local law enforcement (Blackstone and Hakim 2013). Some consumers and businesses desire more police protection than they normally can obtain. In particular, high income communities are the probable demanders of private security services. In the Central Business District of Philadelphia, Pennsylvania private security supplemented public police. In 1991 existing businesses in the cbd requested City Council to impose a permanent five percent surtax on their property taxes to fund private security. Between 1993 and 1994 crime decreased by six percent in the center l city business district but increased by 1 percent in the central police district which includes the center city business district. Further, 78 percent of area population believed that the center city business district was less safe prior to the arrival of the private security guards (Blackstone and Hakim 2010,371). A similar situation occurred in Chicago where residents voted for a special district which involved a supplementary property tax to fund hiring private police. Unlike Philadelphia these officers were armed and acted much like the public police. Indeed 17 percent of their time was spent on serious crime related activities (Blackstone and Hakim 2010, 371). New Orleans has similar such supplementation by private security. In 2012, e almost 30 districts within New Orleans voted to be taxed in order to procure private security services. The state legislature must approve the neighborhood's voting to create such a district. Each resident property is assessed an annual fee, 'usually hundreds of dollars' (McCarthy 2012). One official noted that people are concerned about their security, and want to see more officers available (McCarthy 2012). Some cities say such security districts are inequitable in that the wealthy receive better security. Further, it could be argued that citizens were already paying for protection and alternatively the entire city could add to its police force. One New Orleans district, the Upper Audubon Security district, charges each property owner annually $500, has an annual budget of $200,000 and provides private patrol, personal home escorts, and residence checks. A larger security district, the Mid-City Security District, has a budget of $1 million and the district's president views the districts as a gated community (McCarthy 2012). This arrangement like the others discussed allows residents within a large governmental entity to obtain greater security services than normally would be provided them. Oakland, ca has seen a growth of private security to supplement police services. Wealthy neighborhoods have contracted with private security to patrol their streets. The unusual aspect is the banding together of groups of neighbors to employ private security (Stein 2013). Gated communities have been characterized as a kind of club good where residents band together to purchase collective services for their exclusive use. Included within those services is security. Physical and environmental barriers along with a cohesive community are employed to achieve such security (Csejalvay 2011, 736-7). Even in the 1990s, 2.5 million American families were already living in such gated communities (Blakely and Snyder 1998, 53). Access control is usually a prominent feature of gated communities which originally began in the West and then spread to the East. They usually exist in metropolitan areas and are rare in New England and the deep South. Surveys indicate that security was a pri- mary motivation for living in gated communities. One survey found that 70 percent of gated community residents say that securitywas an important consideration in their decision to live in a gated community (Blakely and Snyder 2011). No surprisingly, income is an important element explaining who lives in such gated communities. The affluent residents are able to obtain more services including security than the less affluent city residents. The gated community allows residents to increase their use of security services. Gated communities are most prevalent in Mexico where in 2010 an estimated 56 million people live in such gated communities of the total urban population of 88 million (see http://en.wikipedia.org/wiki/gated _community). Income differences and the fear of crime encourage such living arrangements. For example, the average 2008 income of Mexican urban residents was $26,654 while rural residents who often live close to urban areas average $8,403. Around the world, gated communities are employed to protect residents from crime, clearly indicating that the residents want more security than provided by the public police. As in Mexico, gated communities with substantial private security are most common in nations with great disparity in income distribution. Examples include Brazil, Saudi Arabia, and South Africa. Volunteer Efforts This category includes neighborhood watch, safety control committees in apartment complexes, citizens serving as auxiliary unarmed police, and safe haven homes. Bennett, Holloway and Farrington (2006) report that in the early 2000s, six percent of uk homes or 27 percent of the population lived in areas covered by neighborhood watch. They note that there were 155,000 neighborhood watch organizations operating at the time. The us had 41 percent of its population living in neighborhood watch covered areas during the early 2000. These volunteer efforts were the largest supplement to public police, and provided information to the police on suspicious activities. The study also stresses the fact that such security alert groups was shown to deter criminals. Self-Protection When public police and local citizen group efforts are insufficient in addressing individual security preferences, self-protection measures are utilized. Residential Self- protection from crime is categorized into deter- ring, prevention, and detection measures (Hakim and Blackstone 1997, 59-60). Deterring efforts are aimed to create the impression that the residence is occupied even when it is not. These measures are designed to encourage the burglar to dismiss the property from consideration when browsing for a target. It includes lights, active appliances, car always on driveway, the absence of accumulated mail and newspapers, and trimmed bushes near windows and doors. Prevention efforts are aimed to slow down or prevent by physical measures the entry of the intruder into the premises. These measures include bars on windows, deadbolt locks, and sash on windows. Detection measures are aimed to alert the police, private security or any pre-assigned person about a possible intrusion. The only such measure is a burglar alarm where a signal is dispatched. Interestingly, a yard sign which signifies the existence of a burglar alarm appears as a significant deterring measure (Hakim and Blackstone 1997,6670). In their empirical study which is based on residents' questionnaires, Hakim and Blackstone (1997, 70) showed that the motive for installing a burglar alarm is mostly for personal security. Property Insurance Insurance policy is a supplement for police aimed at recovering mostly monetary losses resulting from crime. Insurance is a normal good which is positively related to income and wealth. A supplement to public police is the acquisition of insurance policy. An insurance policy will be maintained as long as the expected costs of a break-in are higher than the discounted value of the annual premium payments. In a related matter, Hakim, and Blackstone (1997, 59-75) calculated that insurance discounts offered to owners of burglar alarms are beneficial to insurers. The premiums are beneficial to policy holders considering the costs of the associated treatments resulting from the violent crimes and the deductibles incurred on the property loss. Indeed, Loader (1997) notes the discounted insurance premiums for installing security hardware like burglar alarms, ccTv cameras, and deadbolt locks. Environmental Design Another personal supplement to public police is to restructure the physical layout of Communities to allow residents to control the area around their home. Newman (1972; 1996) pioneered the research and implementation of the defensible space. In his two books, Newman suggested design of streets, the grounds, and access to residents. He also dealt with the design of the lobbies and hallways within housing complexes. His premise was to help people preserve those areas in which they can realize their community held values and lifestyles. The key element in Newman's theory is to create a residential environment where physical characteristics including building layout and site plan function to allow inhabitants to observe their surroundings, and exercise control through effective ownership of their environment. By promoting a sense of'belonging' for the interior and exterior common space, a criminal stands out and feels vulnerable. Newman also observed that smaller multifamily units create greater sense of belonging, better visibility of the environment, and thus make a long term safer living. Conclusions The Tiebout model suggests that a household moving to a metropolitan area chooses among the large number of suburban and urban localities to locate where the mix of public services best reflects its own preferences. The large number of localities provides greater social welfare. The result for a multi-communities region is that demands for public services are likely to be more diverse among than within communities. This paper extends the traditional Tiebout model by considering security services, and suggesting based upon preferences the permanent mixture of public and private security for each community. The theoretical model shows how private security supplements public security and the magnitude varies among localities or preferences. The supplementing of public police with private security is implanted in the five forms of private police, volunteer efforts, self-protection, insurance, and environmental design. All these five forms result at different magnitudes among localities of varying preferences and without government intervention (the invisible hand). The use of private security exists, could increase, and varies among communities even when such services are perfect substitute to public police. The paper investigates three models where quantity demanded for security varies within a community. In the case that the population is homogeneous in wealth, income, and preferences, supplementation is not needed. The society coordinates and shares the burden of optimal pure public good expenditures. This was illustrated by the first model. In the second model where population groups are heterogeneous either by location, properties, incomes, or preferences, a rectangular distribution of the demand for security is generated. In that case the solution of a solely pure public good supply according to the median representative consumer and equally sharing the burden of finance is not the first best solution. By allowing a combination of a pure public good with private types of security, model three may lead under certain conditions to an improvement in social welfare. Several additional implications can be derived from our models. An increase need for security due to objective or subjective factors does not affect the demand for supplemental private security (which can be defined as 'neutrality of private security'), and will be supplied only by additional expenses on pure public security. These demand factors include, among others, changes in property values, income of all population groups, or uncertainty about economic and social conditions. On the other hand, an increase in the production cost of security increases the demand for the pure public good and reduces private security expenses. The reason is that in relative terms, the public good is cheaper and private security becomes more expensive. Thus, the substitution effect permanently dominates in more public security and less private security. The mixture has changed, and in addition the total security level has unequivocally declined. The last important finding is the impact of a change in population. An increase in the population has two effects: On the one hand the larger community yields greater demand for either public or private security. However, there is another effect of the increase in population in our model; a larger community also leads to a larger diversification of demands. These two effects lead to the important conclusion that as population increases public security as well as private security both increase. However, the former increases in smaller percentage terms than the latter. Thus, the ratio of private security to public security increases as the population grows. Appendix 1 dMix dC = 4 = 4 CN - +--C N 2 CN N + -~C CN — +--C N 2 < 0 368 Uriel Spiegel, Tchai Tavor, Simon Hakim, and Erwin A. Blackstone Assuming that N»Awe find that ^ —> o, therefore [-(A-q-?] 4A2 < 0. Make more sense if the attitude towards C is to join mutual financing by customers of pure public security. Appendix 2 dMix + + dN " 4(A-0 or dMi, + f+ § + £) dN " 4 (А-§Г Assuming that N> Cwe find that ^ —> o, therefore dMix _ -C)-A dN ~ 4A2 > 0. Otherwise the sign of is ambiguous. Acknowledgments An earlier version of this paper was presented at the International Conference of the American Real Estate & Urban Economics Association (areuea) in Jerusalem, Israel, 23-26 June 2013. The authors thank the conference referees for their helpful comments. References Bennett, T., K. Hollway, and D. P. Farrington. 2006. 'Does Neighborhood Watch Reduce Crime? A Systematic and Meta-Analysis.' Journal of Experimental Criminology 2 (4): 437-58. Blackstone, E. A., and S. Hakim. 2010. 'Private Policing: Expenses, Evaluation and Future Direction.' In Handbook on the Economics of Crime, edited by B. L. Benson and P. K. Zimmerman, 359-77. Cheltenham: El-gar. --. 2013. 'Competition versus Monopoly in the Provision of Police.' Security Journal 26 (2): 157-79. Blakely, E., and M. G. Snyder. 1998. 'Separate Places: Crime and Security in Gated Communities.' In Reducing Crime through Real Estate Development and Management, edited by M. Felson and R. E. Persen, 53-70. Washington, d c: Urban Land Institute. Buchanan, J. M., 1965. 'An Economic Theory of Clubs.' Economica 32 (125): 1-14. Csejalvay, Z. 2011. 'Gated Communities for Security or Private? A Public Choice Approach and the Case of Budapest.' International Journal of Urban and Regional Research 35 (4): 735-52. Hakim, S., and E. A. Blackstone. 1997. Securing Home and Business: A Guide to the Electronic Security Industry. Boston, ma: ButterworthHeinemann. Holcombe, R. G. 1989. 'The Median Voter in Public Choice Theory.' Public Choice 61 (2): 115-25. Loader, I. 1997. 'Private Security and the Demand for Protection in Contemporary Britain.' Policing and Society 7 (3): 143-62. McCarthy, B. 2012. 'N. O. Residents Increasingly Turning to Private Police Patrols.' http://www. com/news/eyewitness/brenanmccarthy/brenan -maps-176820771.html McCrie, R. D. 1992. 'Three Centuries of Criminal Justice Privatization in the United States.' In Privatizing the United States Justice System, edited by G. Bowman and S. Hakim, 12-26. Jefferson, nc: McFarland. McMillan, M. L. 1989. 'On Measuring Congestion ofLocal Public Goods.' Journal of Urban Economics 26 (2): 131-7. Müller, W. C., and V. Wright. 1994. 'Reshaping the State in Western Europe: The Limits to Retreat.' Wesf European Politics 17 (3): 1-11. Newman, 0.1972. Defensible Space. New York: Macmillan. -. 1996. Creating Defensible Space. Washington, dc: Office of Policy Development and Research, us Department of Housing and Urban Development. Oates, W. E. 1988. 'On the Measurement of Congestion in the Provision of Local Public Good.' Journal of Urban Economics 24 (1): 85-94. Shearing, C., and J. Wood. 2003. 'Governing Security for Common Goods.' International Journal of the Sociology of Law 31 (3): 205-25. Stein, C. 2013. 'As Cities Lay off Police, Frustrated Neighborhoods Turn to Private Cops.' Christian Science Monitor, 5 April. http://www.csmonitor .com/USA/2013/0405/As-cities-lay-off-police-frustrated -neighborhoods-turn-to-private-cops. Tiebout, C. 1956. 'A Pure Theory ofLocal Expenditures.' Journal of Political Economy 64 (5): 416-24. This paper is published under the terms of the Attribution-NonCommercial-NoDerivatives 4.0 International (cc by-nc-nd 4.0) License (http://creativecommons.org/licenses/by-nc-nd/4-0/). Stock Market Consequences of the Suspension of the Central Bank of Nigeria's Governor Ibrahim Mohammed Ahmadu Bello University, Nigeria imohammed@abu.edu.ng Chioma Nwafor University of Glasgow, United Kingdom cnwafor. cn@gmail.com The sudden announcement of the suspension of the Governor of the Central Bank of Nigeria (cbn) on the 20th February 2014 created mixed reactions among analysts and market participants in Nigeria and beyond. The objective of this study is to empirically establish the reaction of listed firms' stock prices to the announcement of the suspension of the Governor of the cbn. Using the standard event study methodology on a sample of 104 out of the 122 listed firms that traded on the floor of the nse on the fateful day, the study sought to establish the significance of abnormal return and cumulative abnormal return on the announcement day, and fifteen trading days after the announcement became public. The study found the presence of statistically significant abnormal return and cumulative abnormal return of -0.06 percent and -5.95 percent on the announcement day. It also established the presence of statistically significant cumulative abnormal return of approximately -6.91 percent fifteen trading days after the announcement. The study concluded that the sudden announcement of the suspension of the Governor of the cbn gave rise to a negative market reaction by listed firms in Nigeria, and the negative trend persisted for the fifteen trading days after the announcement. It was recommended that subsequently, policy makers should as much as possible avoid sudden announcements of the suspension or removal of the Chief Executive Officers (ceos) of public institutions that have close links with the stock market. Where the need for such action becomes inevitable, the announcement should be preceded by the release of information that will minimize asymmetry between policy makers and the stock market. Key Words: stock prices, Governor of the cbn, event studies jel Classification: G12, G14, G21 Introduction The role of the financial system in mobilizing funds from the deficit to the surplus units of any economy cannot be overstressed. The financial sys- Managing Global Transitions 12 (4): 371-394 tem ensures that resources are directed from the surplus spending units to the most productive sectors of the economy. At the centre of this important intermediation function is the stock market, which ensures channelling of resources into long-term productive investments. The stock market of any economy therefore sees to the mobilization of funds on long-term basis to stimulate economic growth. Thus, it is therefore not surprising that a number of studies in developed and emerging markets have empirically documented the role of the stock market in fostering economic growth (Atje and Jovanic 1993; Yartey 2007; Aruwa 2009). A central question to the operation and performance of stock markets all over the world is the extent to which such markets instantaneously and unbiasedly impound new information into stock prices. Thus, a market is considered efficient if it quickly and automatically adjusts to reflect new information. This process of stock prices adjustment to new information is referred to as market efficiency in finance. Fama (1970) identified three forms of efficiency associated with stock markets. According to him, a stock market is said to be weak-form efficient if information on past stock prices is fully reflected in current prices, semi-strong form efficient if all publicly available information is captured in stock prices; and strong-form efficient if all information including the one held by insiders is fully reflected in stock prices. On the other hand, banks complement the functions of the stock market by providing short-term credit and liquidity to the various players within the financial system. Hence, banks also play a very crucial role on the intermediation process by mobilizing resources from the surplus to the deficit spending units albeit on short-term basis. Being profit maximizers, the resource allocation process of banks is highly regulated and often done within the framework provided by the apex bank, which core mandate is to ensure macroeconomic and financial stability. The central bank of any economy therefore plays an indispensable role in promoting the stability of its financial system. Central banks are headed by CEos who are referred to as governors in some economies and presidents in others. According to Lassoued and Attia (2013), ceo attitude can have serious effect on the financial, investment and operational decisions of his/her organization. Thus, as the ceo of a corporate organization, the integrity and independence of the ceo of a central bank plays a crucial role in determining the confidence the general public will have in the economic system locally and internationally. Most of the extant studies on ceo turnover centre exclusively on cor- porate organizations, thereby ignoring public institutions. Previous studies on ceo changes such as Adelegan (2009a) and Bonnier and Bruner (1989) have argued that since ceos and boards of organizations have power of influence over the firm's strategy, policy and decision-making, change of ceo s or board members will be a significant event that could have implications for the firm's market value. This implies that the removal or suspension of the ceo of a sensitive public institution such as the central bank will have market consequences. This is especially true for banks that are most of the time the direct target group of the central bank. Like other value-relevant announcements, stock prices reaction to the announcement of the removal or suspension ceos are studied using the event study approach because of its ability to accurately capture the impact of an event announcement (Ball and Brown 1968). This methodology involves the thorough analysis of the difference between the return earned as a result of the announcement and the return that would have been earned had the announcement not been made (Brown and Warner 1985). The analysis is usually for a defined time period around the date of announcement (event window). The presence of abnormal return at whatever level is an evidence of semi-strong form inefficiency (Peterson 1989). In Nigeria, the sudden announcement of the suspension of the Governor of the cbn on the 20th February 2014 has generated controversy and many reactions among observers and market analysts as to the likely consequences of such an action on the market value of listed firms. It was worrisome to many stakeholders that policy makers could take such an abrupt decision without recourse to the likely consequences such action may have on the Nigerian capital market in particular and the financial system in general. While a number of opinions exist on the extent of listed firms' stock prices reaction to the announcement of the suspension, they can at best be considered as mere conjectures and not products of empirical research. The need to conduct an empirical study to establish the extent of this reaction therefore becomes imperative. Consequently, this study aims at empirically analyzing the reaction of listed firms' stock prices to the announcement of the suspension of the Governor of the cbn. The specific objectives include to: 1. Establish the significance of listed firms' abnormal return on the announcement day of the suspension of the Governor of the cbn. 2. Establish the significance of listed firms' cumulative abnormal return on the announcement day of the suspension of the Governor of the cbn. 3. Establish the significance of listed firms' cumulative abnormal return fifteen trading days after the announcement day of the suspension of the Governor of the cbn. To achieve this, the paper hypothesized that there exists no significant abnormal and cumulative abnormal return within listed firms' event window on the announcement day of the suspension of the Governor of the cbn. The hypotheses are presented in statement and notational forms as follows: H1 There is no significant abnormal return on the announcement day of suspension of the cbn Governor (AR(t0) = 0). H2 There is no significant cumulative abnormal return from fifteen trading days before the announcement day to the announcement day of suspension of the cbn Governor (cAR(t-15,t0) = 0). H3 There is no significant cumulative abnormal return from fifteen trading days before the announcement day to fifteen trading days after the announcement day of suspension of the cbn Governor (CAR(t-ls,t+iS) = 0). The remainder of the paper is structured as follows: section two provides a review of literature and theoretical postulations on stock prices reaction to CEo/board changes and sudden ouster of corporate boards, section three discusses the methodology, section four presents the results and discusses the findings; and section five concludes, draws policy implications, and recommends the appropriate course of action. Literature Review Studies on stock prices reaction to value-relevant announcements are usually conducted by examining the market's response to the disclosure of an event. Identifying previous studies, whether in developed or emerging markets, where the governors (or presidents) of their central banks have been suddenly suspended or sacked becomes very difficult for two reasons. First and foremost, policy makers always avoid such kind of decision because of its likely effect on the economy as a whole, and secondly, the central banks of most economies have clearly established laws and excessive checks regarding the suspension and sack of c eo s of public institutions that are economically and financially sensitive. Thus, the closest link between previous studies and this paper is the study of stock market reaction to changes in ceo s of corporate organizations. Despite the sharp contrast between the operations of corporate organizations and important public institutions such as the cbn, the review of such studies could provide useful insights to understanding the context of this study better. This section therefore presents a review of some studies on stock prices reaction to CEo/board changes and sudden ouster of corporate boards in developed and emerging markets. Adelegan (2009b) investigated the reaction of stock prices of firms listed on the nse to the announcement of change in top management, with a view to establishing whether or not the Nigerian stock market is informationally efficient in that regard. The study employed the traditional event study methodology on a sample of firms listed on the nse from 1997 to 2005. The study documented a significant positive pre-announcement, announcement and post-announcement price reactions. Furthermore, negative stock price reaction was recorded for the announcement of resignation of top management; while the concurrent announcement of resignation, retirement and new appointment of top management gave rise to positive market reactions. The study concluded that top management change in Nigeria is perceived by the market as a positive signal in favour of shareholders' interest. However, a proportion of the scope of the study falls within the period when the nse was not automated, and the study did not to correct for thin trading and volatility effects in the return series. Furthermore, Lassoued and Attia (2013) examined the market effects of ceo turnover in post-revolution Tunisia, using a sample of 16 turnover announcements by 53 firms listed on the Tunisian stock market. The work employed the standard event study methodology in their analysis and found that the announcement of a ceo turnover is on the average bad news for equity investors. The results showed negative abnormal returns following the announcement of ceo change. They concluded that their results are affected by the bear market. A close look at the analysis conducted by the study revealed the absence of unit root, serial correlation test and the test in arch effects as these tests increase the robustness of computed abnormal return. In Nigeria, Osuala, Nto and Akpan (2013) investigated the reaction of the banking sector to the sudden removal of corporate ceo s of some dmbs. Using a sample of five dmbs whose ceos were suddenly sacked by the board of the cbn on the 14th August 2009, the study utilized the event study methodology to report that positive but statistically insignificant abnormal return was observed for the sample banks. The study concluded that the non-significance of the observed positive abnormal return could be explained by the prompt intervention of the cbn through its timely injection of N420 billion in the affected banks. On the contrary, however, the statistical insignificance of abnormal return could be as a result of the violation of the requirement for normality of abnormal return as the evidence of such test was not presented in the work. In a related study, Pessarossi and Weill (2013) analyzed the consequences of ceo turnover on the stock prices of majority of the state-owned listed firms in China. The work employed the standard event study methodology on a sample of 1155 ceo turnover announcements by 658 listed Chinese firms between 2002 and 2010. The study's findings revealed that ceo turnover announcements are associated with positive market reaction for Chinese listed firms. However, the positive reaction is significantly positive only for firms owned by the central government, and not significant for their state and privately owned counterparts. The study concluded that their results provide evidence that ceo turnover in a central state-owned enterprise is an indication of renewed commitment to the economic performance of the firm. Being a relatively volatile market, the study did not take into account the likely effects of volatility on the estimated return. Similarly, Suchard, Singh and Barr (2001) employed a sample of 59 ceo change announcements by 89 out of the 150 listed public firms on the Australian Stock Exchange (asx) between June 1989 and July 1995 to examine the market effects of ceo turnover for Australian firms. Using the standard event study methodology, the study found a negative lagged market reaction on the announcement day of the ceo change. It concluded that the negative market reaction could be explained by two potential effects: theshort-termdamagefromthe ceo dominatesthepotentiallong-term benefits from a change in ceo, and secondly, the news of change in ceo might trigger the release of previously unknown potential problems or managerial behaviour at the firm. The latter explanation closely resembles the circumstances surrounding the suspension of the Governor of the cbn when shortly after the announcement of his suspension, information about a number of financial allegations against him filtered into public domain. However, the results would have been more robust had the study modelled the effect of serial correlation and heteroskedastic effects. Finally, Warner, Watts, and Wruck (1988) utilized a sample of 351 top management changes by 269 firms listed on the New York and American Stock Exchanges (nyse and amex) between 1962 and 1978 to examine the reaction of stock prices to changes in top management. Using the standard event study methodology, the study found no average stock price reaction at announcement of a top management change. However, there is an inverse relationship between the probability of management change and a firm's stock performance. Being one of the oldest studies, the work did not model for the effect of serial correlation and arch effects. In summary, evidence from the studies reviewed on ceo change announcements in developed and emerging markets is overwhelmingly in favour of negative stock prices response to such announcements. It was also seen that some of the studies reviewed had methodological challenges such as the absence ofcorrection for thin trading, serial correlation and arch effects. This study adopted the efficient market theory as the bedrock upon which the analysis rests. The efficient market theory developed by Fama (1970; 1991) holds that in an efficient market, stock prices adjust instantaneously to reflect new information such that it becomes difficult for an individual to trade on such information exclusively. New information could negatively or positively impact on stock prices, depending on the market's judgement of the information. According to the work, managers can communicate to the market about the prospects of firms through information releases. This theory adequately explains the study in that the announcement of the suspension of the cbn Governor may be an attempt to convey to the market the information at the disposal of policy makers in Nigeria. Thus, it will not be out of place to say that stock prices of listed firms in Nigeria (particularly dmbs) will definitely react to the announcement, especially considering the role the cbn plays in achieving financial and economic stability. Methodology The population of this study consists of the 122 listed firms on the Nigerian Stock Exchange that traded on the 20th February 2014. The study utilized all listed firms in Nigeria that traded on the floor of the nse on the 20th February 2014, which was the day the news of suspension became public. For a firm to be part of the sample however, the following criteria must be met: • Data on daily stock prices must be available for the bank at least over the period of 120 trading days before the announcement and another fifteen trading days after the announcement. This criteria resulted in the elimination of three firms. • The bank did not undergo technical suspension within the 120 trading days before the announcement day, and fifteen trading days after it. Six firms could not meet this criterion and were therefore dropped. • No other simultaneous important announcement such as earnings and bonus issues announcement have taken place and contaminated the effect of the event within the event window. The first criterion resulted in the elimination of three firms, whiles the second and third resulted in dropping six and nine firms respectively. Thus, application of the above criteria resulted in a total observation of 104 listed firms. The study utilized the standard event study methodology advocated by Mackinlay (1997) where abnormal return is computed as the prediction errors of the market model. The methodology have been found over time to be consistent and valid in measuring the impact of important corporate events such as stock splits, bonus issues, mergers and ceo sudden ouster (Bonnier and Bruner 1989; Shaheen 2006). In this study, the event is defined as the announcement of the suspension of Sanusi Lamido Sanusi as the Governor of the cbn. This study treats the announcement of the suspension as technically synonymous to removing the Governor because the two events are likely to have very similar market effects since they both suggest discontinuity in the policies and programmes of the suspended or ousted ceo. An event window of thirty-one trading days, covering fifteen trading days prior to the announcement day, the announcement day itself, and another fifteen trading days after the announcement day was utilized by the study. Although Panayides and Gong (2002) opined that an event window of 11 trading days covering five days before the announcement and five days after it is sufficient to fully capture the effect of an event, the study utilizes a larger event window because emerging markets like Nigeria are generally known to be less efficient than matured markets and thus tend to be more sluggish in reflecting new information in stock prices (Afego 2010). It is therefore expedient to choose a reasonably large event window to accommodate this speed of adjustment. In line with De Medeiros and Matsumoto (2006), an event window of thirty-two trading days, covering fifteen trading days before the announcement, the two-day announcement date and another fifteen trading days after the announcement was utilized by the study. Furthermore, the study also utilized a parameter estimation window of one hundred and twenty trading days (from day -16 to day -135) over which the parameters for normal return were estimated. According to Dyckman, Philbrick and Stephan (1984), Brown and Warner (1985) and Shaheen (2006), a parameter estimation period of 120 days is adequate since daily return data for the 120 days prior to the event can sufficiently formulate a benchmark for normal returns. This study relied solely on data collected from secondary sources. Specifically, the study utilized secondary data relating to the daily stock prices of sample-listed firms for the period under study. Similarly, the corresponding nse daily All Share Index (asi) was collected for the same period. Both the daily series of stock prices of the sample firms and the corresponding nse asi were retrieved electronically from the online database of Cashcraft Asset Management Limited. The daily stock price data collected was then converted to daily continuously compounded stock return. Given that the study utilized an event window and estimation window of less than a year, dividends were constant at zero. The logarithmic transformation of the time series data became necessary in view of the need to keep the effect of outliers under control. The same approach was also applied to the nse asi to create daily continuously compounded market return The log daily returns of sample firms and those of the nse asi were the main variables employed to estimate the model for generating individual bank and market returns. Being time series data, the daily stock and market return series were subjected to stationarity test using the Augmented Dickey-Fuller (adf) test for the presence of unit root. Although the daily firm and market returns were computed using the market model, they are not free from inherent statistical bias due to the effect of thin or infrequent trading. According to Abuzarour (2005) and So-hawon (2006), emerging markets like Nigeria are typically characterised by low liquidity and thin trading. Furthermore, the study by Tijjani et al. (2009) clearly revealed the presence of significant thin trading in the Nigerian stock market, especially the petroleum and banking sectors. Thus, given that the observed index in thinly traded markets does not represent the true underlying index value, there is always a systematic bias towards rejecting the efficient market hypothesis (emh). Against this backdrop, this paper corrected for the effect of thin trading in the data. The study employed the correction procedure introduced by Miller, Muthuswamy, and Whaley (1994). According to them, thin trading correction reduces the negative correlation among returns. The methodology proposed by Miller, Muthuswamy, and Whaley (1994) suggests that a moving average model (ma) that reflects the actual number of non-trading days should be estimated and then returns adjusted accordingly. However, given the difficulty in identifying the actual non-trading days, Miller, Muthuswamy, and Whaley (1994) have shown that it is similar to estimating an ar(1) model from which the trading adjustment can be obtained. The model, as advanced by Miller, Muthuswamy, and Wha-ley (1994), involves estimating the following equation: Rt = ai + a2Rt-i + St, (1) where a1 and a2 are parameters to be estimated (a1 is the slope and a2 is the coefficient of the ar(1) term), Rt is the index return at time t, Rt-1 is the index return at time t - 1, and st is a random disturbance term. Using the residuals from the regression, adjusted returns were estimated as follows: Rf = (2) 1 1 - a2 where is the return at time t adjusted for thin trading, and st and a2 are as defined above. The model above assumes that the adjustment for non-trading is constant overtime. In order to capture the abnormal returns in the event window, the study utilized the market model pioneered by Ball and Brown (1968), Fama et al. (1969) and Brown and Warner (1985). The single factor market model was employed to estimate the return within the event window and the parameter estimation window. The model is presented as follows: Ri,t = a + ßRm,t + St, (3) where Riit is the return on firm i at time t, a and ß are parameters to be estimated, Rm,t is the return on the market index at time t, and st is the stochastic error term, which is random and follows a normal distribution pattern. To ensure robustness of the estimated residuals from the model, equation (3) was estimated using the Newey-West's (Newey and West 1987) Heteroskedasticity and Autocorrelation-Consistent (hac) estimator, which automatically corrects for autocorrelation and heteroskedastic effects in residuals. However, these were still checked for in the residuals of all the estimated models. Assuming a constant beta value, the estimated return for firm i's security can be computed by substituting the estimated values of ai and ßi over the estimation window in equation (3) above as follows: Eitt = "a i + ßiRm,t, (4) where E;>, is the expected return on bank is security at time t; a) and ßj are the estimated parameters based on the estimation window; and Rm,t is the market return at time t. The abnormal return is defined as the difference between equation (3) and equation (4) as follows: ar = Ri>t - Ej,t. (5) Once the estimated equation has been obtained, the actual return on firm is security is calculated as follows: Ritt = ßi + ßiRm,t + Ei,t. (6) Since Ец = a) + ßjRm>t equation (6) simplifies to: Ri,t = Em,tSi,t. () This implies that abnormal return for firm i at time t is simply given as: AR = Sit (8) Thus, the abnormal return on the security of a given sample firm is simply the residual of the ols after regressing the firm's stock return on the market return. For the residuals to be considered as the abnormal return however, the parameters estimated over the estimation window must be integrated into the equation as shown above. Although Brown and Warner (1985) have concluded that estimates from ols using the market model are generally well specified and display no significant mean bias, it has been empirically documented that the ols estimation fails to adequately capture arch effects in returns series. According to De Medeiros and Matsumoto (2006) and Brooks (2008), estimating a model that adequately captures arch effects is important because their existence renders the coefficient estimates inefficient and thus produces a downward bias in abnormal returns. Studies by Akgiray (1989), Bollerslev (1986) and Chiang and Doong (2001) have shown that daily stock returns exhibit variable volatility along time, thus tending to show arch effects. Similarly, the study by Emenike (2010) concluded that the Nigerian stock market index return exhibits significant volatility and the presence of arch innovations. Consequently, this study employed the Engle (1982) test to check for the presence of arch effects in the residuals of the market model over the parameter estimation window and the event window. In the event that significant arch effects were detected, the Öls market model estimation for the affected sample firms was re-estimated using arch or garch models according to their best fits. When a garch (1,1) model is considered, equation (6) is replaced with: = aio + ahu2i>t-1 + ah