«Working Paper No. 77 January 2011 b B A M B AMBERG E CONOMIC R ESEARCH GROUP k k* BERG Working Paper Series on Government and Growth Bamberg Economic ...»
Political cycles in income from privatization
The case of Albania
Working Paper No. 77
BERG Working Paper Series
on Government and Growth
Bamberg Economic Research Group on Government and Growth Bamberg University Feldkirchenstraße 21 D-96045 Bamberg Telefax: (0951) 863 5547 Telephone: (0951) 863 2547 E-mail: firstname.lastname@example.org http://www.uni-bamberg.de/vwl-fiwi/forschung/berg/ ISBN 978-3-931052-87-4 Reihenherausgeber: BERG Heinz-Dieter Wenzel Redaktion Felix Stübben∗ ∗ email@example.com
POLITICAL CYCLES IN INCOME FROM PRIVATIZATION
THE CASE OF ALBANIA
DRINI IMAMIFaculty of Economics and Agribusiness, Agriculture University of Tirana firstname.lastname@example.org
ENDRIT LAMIPhD student, Budapest University of Technology and Economics email@example.com
Earlier research on PBC in Albania found clear evidence of fiscal expansion before elections. In addition to increased income from taxes and borrowing, another source of financing the increased fiscal expansion in transition countries may be income from privatization, which is also the object of the analysis of this paper. In our analysis we apply standard econometric approach, used widely for research related to PBC.
We test if income from privatization increases before elections.
We find statistically significant increase of income from privatization before general (parliamentary) elections, which may lead us to conclude that one of the reasons may be to finance increased expenditures before elections. Another motivation, behind this behavior of the incumbent, may be rent – seeking.
These results are of particular interest, as it is for the first time that income from privatization is analyzed in conjunction with PBC.
Keywords: Albania, Political Business Cycle, Privatization JEL Classification: E32, O23, H61 Political Cycles in Income from Privatization
1. Introduction Sometimes it is called “Election years economics”: The long run interest of incumbents are largely dominated by the view on the next election, hence, they start to cycle around the question how to be reelected as a wolf does with the pray. There is a common sense that the economic performance before elections, determines, to a large extent the likelihood of reelection for the incumbent, and the other way around (Tibbits, 1931). Hence, economic factors influence political factors but also the political factors may influence economic ones – governments may use all means they possess, including economic policy instruments, to enhance the chances of reelections. Therefore governments may engage in expansionary economic policies prior elections, increasing output thereby trying to decreasing unemployment, in order to please the voters, creating this way Political Business Cycles (PBC).
Over the last decades, there has been plenty of research and publications on PBC, aiming for analyzing and explaining the use of fiscal and monetary instruments by the incumbent to stimulate economic performance before elections. The government may behave opportunistically and inefficiently prior to elections engaging in expansionary economic policies, to increase the output and decrease unemployment, in order to please the voters, however, creating Political Budget Cycles and maybe also Political Business Cycles. The PBC model of Nordhaus (1975) opened the way for many following empirical and theoretical studies and publications and remains a point of reference. While initially, the focus of PBC related empirical research was focused on Western Countries, over the last decade there is growing interest of research on PBC in developing and/or transition countries, whose institutions, economies and societies differ significantly from those of developed/western countries. As shown by Brender and Dazen (2005) and Shi and Svensson (2006), new democracies are particularly vulnerable for political budget cycles. While Alt and Lassen (2006) show the relevance of transparency, Brender and Dazen (2005) also pronounce the lack of experience that voters have in new democracies regarding the existence of political fiscal cycles. Meanwhile, Shi and Svensson (2006) see beside the aspect of information asymmetry also the incumbents’ rents out of staying in power as a relevant issue.
Evidence of PBC was also found in several less developed and new democratic countries.
Gimpelsen (2001) proved the existence of PBC in Russia, and Asutay (2004) provided evidence for the presence of PBC in Turkey. Meanwhile Hallenberg and Souza (2002) prove the existence of PBC of EU Accession Countries, in both forms, fiscal and monetary instruments (the later, was more common in countries with low level of independence of the Central Banks).
Recently, there has been growing interest for research on PBC in Albania, also looking into phenomena typical for transition economies (Imami and Lami, 2006; Imami and Lami 2008;
Lami, Imami and Kächelein 2008; Kächelein, Imami and Lami 2008). According to this previous research, there is clear evidence of Political Budget Cycles, namely increased public investments before elections (Imami and Lami, 2006). One of the sources of financing the state budget in transition countries, characterized by a massive privatization process, may be income from privatization. That can also be the case if expenditures increase in conjunction with elections. In this context, our hypothesis is that the incumbent may engage more intensively in privatizations before elections, aiming at increasing public revenues to sustain increased expenditures, in the context of Political Budget Cycles. Hence, this paper focuses less on the existence of Political Budget Cycles or Political Business Cycles per se, but rather on incumbents potentially use of public assets through privatization to finance such election related cycles.
In the upcoming chapter we will provide a short overview of the privatization process in Albania as a background of the hypothesis. Chapter three provides an overview about the method and data used while chapter four presents the main findings and chapter five concludes the paper.
2. Searching for cycles in income from privatization
2.1 Background of the hypothesis In Albania, the privatization process started in the early 1990’, in the context of the transition from planned economy to market economy. Large-scale privatization started in 1992, as part of a privatization program guided by the IMF and the World Bank, whose result was that the economy was largely in private hands by 1996 (MIGA, 2002). The privatization process was put in halt during the unrest of 1997, but resumed again in the following years. Based on the EBRD index, small scale privatization can be seen as finalized since 1995; meanwhile large scale privatization persisted on a moderate level until end of the last century (EBRD 2004: p.92, 2005: p. 96).
During 2000’, there were privatized the banking sector (currently all secondary level banks are in private hands), Albtelecom (the landline monopoly company), OSSH (the Albanian electricity distributor), while the government is committed to sell its remaining share in Albtelecom and OSSH, and also intends to privatize INSIG a leading insurance company (Prifti 2010). Again based on the EBRD index, large scale privatization improved mainly in the last years as 2008-2010. (EBRD 2009: p.134, 2010: p. 4) Privatization processes and agencies are perceived as highly corrupted in Albania (Muço, 2000). There are indications, that privatization is often used in Albania (but also in other transition countries) to finance increased expenditures, also in conjunction to elections. Some of the decisions on (partial or full) privatizations of key publicly owned company, as telecommunications and energy, were taken months before elections. This phenomenon has taken place under different governments, and often such decisions were deemed by the opposition, media and economists as related to the elections.1 In this paper, we test, for the hypothesis, that the incumbent engages more intensively in privatizations before elections, aiming at increasing public revenues to finance increased public expenditures before elections. Raising revenues as a motivation for privatization is widely discussed in the related literature of privatization. Vickers and Yarrow (1991, p.118-19) pronounce the motivation of less developed countries to privatize for revenue purposes as they
As a tentative background information see for example Alsat (2009), Gazeta Shqiptare (2008), Mitrovicapress (2008) and VoA (2005).
may be restraint on the bond market due to inflation risks. Nevertheless, this argument focuses mainly on the question of market access to bonds. Politicians are also constraint in other ways as public perception and restraints placed by other lenders (e.g. IMF).
3. Method and Data
3.1 Specifications of Variables, Data and empirical tests The time series of income from privatization is monthly, spanning from M1-1999 to M10from January 1999 to October 2009), adding up to 130 observations. There are three parliamentary elections taking place in this period, namely June 24, 2001, July 3, 2005 and June 28, 2009.
Following the standard approach in this area,2 we will apply the Intervention Analysis based on Box and Tiao (1975), a methodology for constructing a statistical model in our study. In this paper we test the hypothesis of the existence of changes in the income from privatization before elections. Basically the test proceeds by subjecting the monthly seasonally adjusted time series of this variable to a Box-Tiao intervention analysis using the most appropriate autoregressivemoving average (ARIMA) for the social process and an intervention term; here the intervention term models the time distance to the election day.
A simple formal representation of the intervention analysis is:
where z denotes the revenues out of privatization, modelled using a suitable ARMA(p,q) model and PDt a political dummy variable specified later on. The parameter ω0 measures the change caused by the intervention as modelled by the political dummy variable and is estimated along with the ARIMA time series component. The estimation procedure provides an estimate of ω0 and a confidence interval for the parameter. We have created two kinds of political dummy variables to capture the impact of the elections on privatization revenues, namely cumulative dummy and discrete dummy.
See for example McCallum (1978), Hibbs (1977), Alesina and Sachs (1988), Alesina and Roubini (1992). Hibbs (1987) offers a good introduction to the Box-Tiao technique.
We have 12 cumulative t = −8, −7,..., −1,1, 2,..., 4 election political dummies ( PDt ) and each
of them is defined as:
In the same manner we defined eight discrete elections dummy variables, covering only the quarterly and not the cumulative effect of the two years before the election and the same for four discrete post election dummy variables. As in case of the first and the last event, the election day was at the end of the quarter, the quarter was counted as prior, while in case of the second it is take as after election.
3.2 Estimation of the empirical model In the first stage, we have followed precisely the Box-Jenkins (BJ) Methodology (1970). In the beginning of the process, the first step was removing the seasonal patterns form the time series. Next we carefully investigate on the stationary of the time series as a necessity in further steps. The original time series of privatization receipts results stationary based on the conventional tests (ACF, PACF correlogram and Augmented Dickey Fuller test). Based on BoxTiao’s (1975) intervention analysis, after ensuring for the stationarity, the time-series is modeled as ARMA (Auto-Regressive Moving Averages). By modeling through ARMA it is possible to prove if elections can explain the changes taking place in household expenditures, in addition to the past history of the variable and the random error term. Hence, it is necessary the identification of ARMA (p,q) benchmark model. To find the “best” ARMA model for each time series we are straightforwardly based on Box-Jenkins methodology (1970). Hence, in order to model the analyzed time series as an ARMA we went thought an iterative process of identification, estimation and diagnostic checking of several ARMA models until we found the most plausible one, deemed as the “best” for each series3. The most appropriate model tentatively found was the one with just only one moving average term with a lag four.