Social activist and Anna Hazare aide Arvind Kejriwal recently focused India's attention once again on the unsavory backgrounds of many of our politicians: “In this Parliament, 163 members have cases of heinous offenses against them…. rapists are sitting, murderers and looters are sitting… How can you expect that you can get reprieve from poverty and corruption,” he asked rhetorically.
I recently wrote about research which tried to understand why alleged criminals are fielded by mainstream political parties. The bottom line is that such candidates often command money and muscle power which can help swing a tightly contested election their way.
They may also be seen by voters as being able to deliver the “goods” (for example, public spending.) As Milan Vaishnav, a PhD. candidate in political science at Columbia University, who studies Indian elections, writes: “Voters are not ignorant or uninformed; they’re simply looking for candidates who can best fill a perceived representational vacuum.”
So maybe it makes sense for people to vote in politicians who are facing criminal charges. But what do these good folks, once elected, deliver to their constituents?
New research by Matthieu Chemin, an economics professor at Canada’s McGill University, asks this very question. Mr. Chemin shared this important research with me in advance of its publication in the Journal of Law and Economics.
There’s certainly a widespread belief in India that allegedly criminal politicians worsen poverty and foster further criminality and corruption.
But what causes what?
As Mr. Chemin rightly points out, those with a criminal bent may thrive in regions where poverty or criminality are worse, rather than bringing about those outcomes. Criminality in politics and poverty are certainly related, but does the first cause the second? Equally, there’s the possibility that both criminality in politics and poverty could be simultaneously caused by some underlying feature of a region, such as historical deprivation. Just observing that criminality and poverty go together doesn’t by itself tell us anything much.
Mr. Chemin finds a clever way around this problem, by exploiting the fact that allegedly criminal politicians often win tightly contested races, as I noted previously. Using a nifty statistical technique called “discontinuity” analysis, he compares constituencies in the central and state elections fought in 2004 to identify those places where allegedly criminal and non-criminal candidates were expected to receive approximately equal vote shares: these were the tightest of tight races. The result of a very tight election is basically determined by a coin toss: a few more people supporting one candidate rather than another show up that day to vote, for instance. This small source of random variation in otherwise tight races is what Mr. Chemin’s study is able to exploit in trying to figure out if allegedly criminal politicians actually cause bad outcomes such as poverty.
So now you’ve voted your local don into office, what are the effects? Mr. Chemin’s headline finding is that allegedly criminal politicians reduce the monthly per capita spending of people in vulnerable groups, in particular, the Scheduled Castes , Scheduled Tribes, and Other Backward Classes. This effect isn’t small: the election of an allegedly criminal politician decreases the monthly spending of the average Scheduled Castes, Scheduled Tribes and Other Backward Classes by a whopping 19%.
Why? Mr. Chemin tests for a straightforward hypothesis: allegedly criminal politicians engage in more heinous acts, such as murder and arson, that occur largely at the expense of these vulnerable groups. Presumably, they’re transferring resources to their favored constituents at the expense of the weakest. The evidence supports this: murders, riots, and arson increased by 19%, 23%, and 34%, respectively, after 2004 in districts which elected allegedly criminal politicians after tightly fought elections. Crimes against women also worsened. Interestingly, there’s no evidence that crimes against property (such as robbery, burglary, and theft — arson is classified as a crime against public order) or other economic crimes (such as breach of trust or counterfeiting) increased.
These results tally with common sense. But Mr. Chemin’s most unusual finding, which might surprise most readers (and certainly members of Team Anna) is that the election of allegedly criminal politicians actually reduced corruption by local bureaucrats. Corruption obviously cannot be easily measured, so the study uses the value of gifts received by bureaucrats as one signal of how corrupt they might be. These dropped by a massive 66% after 2004 in districts where allegedly criminal politicians had won tight races.
But if you think about it for a moment, not really. It’s in the interest of politicians to rein in corruption by bureaucrats, since presumably they want the biggest slice of the illicit pie for themselves. Also, bureaucrats might simply be less relevant and therefore less “bribable” in districts which are run like fiefdoms. Allegedly criminal politicians could also cajole and bully bureaucrats into doing what they want, again making it less necessary to bribe them. This is a situation where a presumed reduction in corruption is bad news rather than good news, since it’s replaced by stuff that may be even worse. Let’s not forget as well that it was accompanied by reduced consumption of the poorest and an increase in criminality.
All of this recent research, capped by Mr. Chemin’s important new paper, raises as many questions as it answers. Why do people keep voting in politicians that face criminal charges if they have such bad effects on the districts they are elected in? It might be the absence of a better alternative, or the fact that, for good or ill, allegedly criminal politicians do deliver something to the folks who vote for them – perhaps a quicker response at a time of need than the local bureaucracy. But, as we saw, whatever benefits they bring to favored constituents come at a high social cost in terms of decreased consumption by the poor and increased criminality all around.
Rupa Subramanya writes Economics Journal for India Real Time.