The aim of this paper is to prove the occurrence of policy tools manipulation by incumbent politicians prior to an election, followed by the reverse policies after the elections are over, in a cyclical pattern. The paper studies the case of post-communist Romania. Our approach uses foreign currencies exchange rates as indicators of a change in the monetary mass, opposed to the price of gold as an indicator of the real economy. The interest rate for deposits is also analyzed through auto-regressive time series. Statistical analysis of data provides clear evidence of political intrusion in economy in Romania.
ARIMA models, deposit interest rate, foreign currency, political business cycle.
The expression of freedom in economy could be interpreted as the lack of political intrusion. Although highly controversial in the academic world, the political business cycle (PBC) is a reality, more obvious in developing countries, including those from Latin America, Sub-Saharan Africa and the former Soviet Union.
Manipulation of fiscal, monetary and legal instruments diverges from the investment behaviour of incumbent politicians, who should safeguard both present and future welfare and becomes an expression of their opportunism.
The succession of political parties with different ideologies could also result in a cyclical pattern in the economic development of the country, but while this is a natural consequence of the voters’ choice and reflects their desires, the dark side of this phenomenon is the artificial expansion of the economy prior to elections, followed by deep contraction afterwards.
The aim of this paper is to bring empiric evidence of the existence of an opportunistic political business cycle in post-communist Romania, between 1995 and 2012. Foreign currency exchange rates for U.S. Dollars and Euro expressed in RON are used to study the change in the monetary mass deployed just prior to elections, in order to simulate an expansion in the domestic economy while the price of gold in RON offers a dimension of the real economic situation, because gold reserves have a lot less liquidity. The interest rate for deposits is also under investigation due to high values prior to elections.
2. LITERATURE REVIEW
Although the term „political business cycle” was coined in 1943 by Kalecki , it did not have an academic resonance until the seventh decade of the twentieth century trough the seminal work of Nordhaus , Lindbeck  and Tufte , which advocated for a pure opportunistic model consisting of greedy politicians and a naive electorate, more formally known as ”an adaptive expectations electorate”. These early approaches involved studies of the inflation and the unemployment rates using the Phillips curve.
The political innocence of the electorate is denied by Hibbs  who introduces the ”partisan elections” hypothesis. He gives great importance to the ideology of the parties engaged in the electoral race, and empowers voters with knowledge on partisan differences. Hibbs’ later work, namely ,  and that of Sheffrin  do not contradict the Nordhaus model, as Roman et al.  notice, under the assumption that politicians may display both opportunistic and partisan behaviour simultaneously.
The early models imply the lack of rationality of the voters, an assumption which is uncommon in economics. As long as we consider voters to be the same players that act on the economic market it is only natural to assume that they use, in this particular decisional process, the rationality which is an intrinsic characteristic. Thus, the natural model that emerges is that of rational opportunism, promoted by Rogoff & Sibert  by including competencies, further developed in , using the informational asymmetry between voters and politicians. Persson & Tabellini  chose to adapt Nordhaus’ model by including a utility function for the voter. More recently, Stein &Streb , considered an opportunistic model of politicians deceiving the electorate just prior to elections using artificial changes in the monetary policy, thus influencing the inflation rate.
The most sophisticated models are the “rational partisan models” belonging to Alesina & Sachs  and further developed by Alesina et al. , which use a utility function adjusted to the ideology of the party. These models often contradict Nordhaus’ model, depending on the political orientation.
Inflation and unemployment rates were the first variables studied in the attempt to illustrate the political business cycle, with satisfactory results for at least two decades, mostly in developed countries.
The case of developing (emerging) countries needs a new approach, new tools to photograph the same phenomenon. Hence the use of the foreign currency exchange rates employed by certain papers: , ,  and  or other monetary policy tools such as risk ratings, bond spreads , , public savings  and fiscal performance . These innovative approaches produced convincing results also in the case of the U.S., as the following papers imply: ,  and .
Romania is a relatively young democracy struggling between an uneducated electorate, and opportunistic politicians. Thus, one appropriate model for this country is an adaptation of Nordhaus’, such as that of Roman et el. . A more regional approach can be found in Jula& Jula .
3. EMPIRICAL RESULTS FOR ROMANIA
A scarce literature on the subject of the political business cycle in Romania gives the opportunity of a new type of analysis, using the foreign currency rates as an indicator. Also, the price of gold is considered as depicting the real economy, because, while the monetary mass can be easily changed, thus influencing the foreign currency exchange rates, the gold reserves and their price tend to have a more stable nature.
As suggested by the Austrian business cycle theory , the interest rate is a key ingredient of an economical pattern, strongly influencing both present and more important, future welfare. Hence, the monthly interest rate for deposits  was included in the list of variables, in order to see if there is a political influence, prior to elections.
The empiric study uses monthly records of four variables: the exchange rate for the US Dollar (USD) and the Euro (EUR), the price of gold (XAU) expressed in Romanian new lei (RON) and the interest rate for population’s deposits (deposit %). The series spread over 16 years (Jan. 1995- Jan. 2012) for USD and XAU, over 12 years (Jan.1999- Jan. 2012) for the EUR and 9 years (Jan. 2003- Jan. 2012) for the interest rate. The foreign exchange rates were retrieved online from Forex Trading through Oanda Corporation  and represent monthly averages of the midpoint of the prices. The quotations were converted in RON, for a uniform treatment, even if data prior to July 2005 was expressed in Romanian lei (ROL).
Two dummy variables were introduced accounting for the pre-election time: politic_6 and politic_12. These variables are 1 if the month considered is at most six (twelve) months before the elections and 0 otherwise. Table 1 illustrates the pre-election months, ideology of ruling party and ideology of the opposition party. We follow the methodology described in .
Time series analysis was employed, namely an autoregressive integrated moving average (ARIMA) model for each of the variables. ARIMA models are used only for stationary series.
Tabel 1-Elections in Romania 1995-2012
illustration not visible in this excerpt
Source: after http://ro.wikipedia.org/wiki/Alegeri_%C3%AEn_Rom%C3%A2nia
3.1 Gold- the expression of the real economy
Although gold is no longer the standard, it is a viable indicator of the economy’s current state. Gold reserves are a strategic patrimony for every country and while precious metals are traded at stock markets like many other commodities, their price is the reflection of a certain country’s real economic welfare. Incumbent politicians try to mislead naïve voters with information about economy closer to their understanding, while the quotation for precious metals is a domain for connoisseurs.
This empiric study proposes a ARIMA model for the gold quotations. Two dummy variables (politic_6 and politic_12) were used to check if the price of gold is influenced by pre-election periods.
The model proposed by using SPSS 17 Create models facility was ARIMA (0,1,0), for the gold price series after stationarization using natural log,with the following parameters: