Bangladesh  has turned the political business cycle phenomenon upside down.
Political business cycles are cycles in macroeconomic variables – output, unemployment, inflation – induced by the electoral cycle. This type of business cycle results primarily from the manipulation of policy tools by incumbent politicians hoping to stimulate the economy just prior to an election and thereby improve their reelection chances.
Expansionary monetary and fiscal policies have politically palatable consequences in the short run. When pursued to excess, these very policies can also have very unpleasant consequences in the longer term in the form of accelerating inflation, decreasing savings, worsening foreign trade balance, and long-term expansion of government's share of the GDP at the expense of private consumption and investment. So immediately after the election, politicians tend to “bite the bullet” and reverse course by raising taxes, cutting spending, slowing the growth of the money supply, and allowing interest rates to rise. As a result, the regular holding of elections tends to produce a boom-and-bust pattern in the economy because of the on-again-off-again pattern of government stimulus and restraint to induce an artificial boom at every election time.
Bangladesh’s experience also shows the existence of a political business cycle in GDP growth, albeit with exactly the opposite pattern of boom and bust. GDP growth has consistently declined in each of the last five election years. It happened in 1991, 1996, 2002, 2007 (an election year without election) and 2009 (Figure 1). From the perspective of Western political business cycle theory these growth tendencies appear suicidal for the incumbent. Instead of expanding the economy faster to gain votes, the incumbents appear to be shooting themselves in the foot by allowing the pace of expansion to slow in the election year!
Is this another case of the Bangladesh paradox?
Not really. It has a fairly straightforward explanation. With the exception of the most recent past election, every election came after a period of political impasse on the issue of how election will be held. Every election year is marked by uncertainty regarding the holding of elections and the peaceful transfer of power, street agitation, strikes and violence as the incumbent and the opposition bargain to press their stance. The consequence for the rest of the populace—households and business—is a period of uncertainty that increase the value of waiting in making significant economic decisions. Consumer confidence and animal spirits drain under such circumstances. This depresses demand and slows capacity creation while the strikes and violence disrupt capacity utilization, thus reducing growth.
One apparent empirical anomaly in the above argument is the absence of violence preceding the FY09 election, yet GDP growth declined. However, there was on-again off-again political impasse in the run up to the election. There is thus a clear perverse relation between election and growth even when election is not preceded by violence. Also, although FY07 was an election year, the election had to be postponed. The impasse and violence related to election due in FY07 happened in the last half of 2006 and consequently growth in FY07 was lower than growth in FY06.
Is it the case that the uncertainty surrounding the political transition is the main culprit? Looking at Figure 2, one cannot be too confident because the decline in growth in election years does not seem to have been driven by decline in investment rates. Excepting FY91 election, investment rate has been higher than the previous year in the other four election years. This implies that elections do not undermine, and may be even inspire, confidence in the post-election state of the economy. Disruptions to capacity utilization appears to have been the main reason for growth decline, although the decline in FY09 election year is not consistent with this because there were no strikes and violence in the run up to the FY09 election. Both uncertainty surrounding the peaceful transfer of power and disruptions caused by strikes and violence seem to matter.