Monday, December 24, 2012

Can economics help us understand violence against women in India?

Can economics help us understand violence against women in India?

Vivek Dehejia

In the aftermath of the Delhi rape case, and amidst the welter of moral outrage, hand-wringing, and public demonstrations, that, understandably, have followed in its wake, it’s important to take a step or two back and ask some deeper questions.

What is the root cause of violence against women? Is it misogyny that is deeply engrained in our still patriarchal culture? Is it the churning of a society in flux, as “old” India confronts “new” India, as some commentators have suggested? Does it have to do with a broken criminal justice system that fails to deliver justice? Is it a backlash against the increasing economic and social empowerment of women?

One of the basic lessons of social science is that a complex phenomenon such as violence against women rarely, if ever, has a unique explanation, a single driving causal mechanism. Rather, such phenomena are “over-determined”, plausibly explainable by a multitude of different hypotheses, which aren’t necessarily mutually exclusive nor exhaustive. In simple terms, all of the factors noted above are likely to be at work, and recent research bears on several of them.

In our new book, “Indianomix: Making Sense of Modern India”, my co-author and I devote a significant portion of one chapter to looking more deeply at the issue of violence against women and probing what the research has to say. The following is based, in part, on the research that we present in that chapter.

First off, there’s much, and legitimate, criticism of the quality of data that scholars have to work with. Rape and other violent crimes against women tend to go under-reported in India, as they do elsewhere, and this certainly “contaminates” (in a statistical sense) the quality of the data. Statistically, the correct response is not to discard such data as worthless, as some have suggested, but to “treat” (again, in a statistical sense) the data appropriately. For instance, in the face of systematic under-reporting of violence against women, statistics on the incidence of rape at any given point of time may be misleading; but its rate of change over time is still statistically meaningful. Thus, the fact that incidence of rape and violence against women show an increasing trend is worrisome.

On the other hand, though, an increase in the incidence of violence against women may also reflect increased reporting of such crimes and a more accurate classification – that is, a crime that was previously categorized as generic violence is now correctly categorized as, for instance, a rape. Some recent research, indeed, suggests that increasing political and economic empowerment of women has had exactly these effects, and so part of the increased reported incidence may actually reflect better reporting rather than an increase in the actual incidence. However, it remains for future research to parse the increase in reported crimes against women into these two components.

Another question that goes unasked in the predictable knee jerk reaction to a tragedy such as the recent one is to probe the deeper causes of the high incidence of violent crime, especially violence against women, in India. In the book, we explore one important hypothesis that comes out of academic research both in India and in other places such as China: that there is a robust statistical correlation between the skewed sex ratio (that is, more men than women) and the incidence of murder and other violent crimes including rape. By no means is it claimed that the sex ratio is the only variable that can explain crime; rather, it is one of a group of other important socio-economic factors.

The great difficulty with any statistical correlation is that it doesn’t necessarily imply a causal relationship, nor does it tell us the direction of causality, in case such a relationship exists. Does an adverse sex ratio cause violence? This would be if, for example, lots of single angry young men are roaming around and are violence-prone as a consequence. It’s noteworthy that in the Delhi rape case, the oldest of the alleged perpetrators is only 33 and the youngest is a juvenile and all but one of them is single. Of course, a single data point doesn’t confirm or refute a theory, but it certainly fits a suggestive pattern. It’s not just India where a skewed sex ratio is seen as an underlying driver of violence including against women. Some scholarship suggests that this explains the increase in violent crime in China, where the sex ratio is even worse than in India.

But in economics, things are rarely so simple. It could equally be that, in a society which already has a skewed sex ratio, families will prefer to have boys rather than girls, both because boys offer greater protection to parents when they grow up and because daughters will be at greater risk in a violent society. A statistical correlation by itself cannot discriminate between these two competing explanations.

What’s more, in the rarefied world of statistics, there are two other possibilities: first, that the correlation is “spurious”, and, second, that there is an “omitted variable” which is the underlying cause of the correlation. In the case of the skewed sex ratio and violent crime, one can safely discard the possibility of spurious correlation, since there are compelling causal mechanisms one can point to running in either direction. But one cannot dismiss the idea that a deeper force is at work which leads both to a skewed sex ratio and more violent crime. The natural candidates are patriarchy and misogyny that are deeply embedded in our culture and have been slow to change. Indeed, it is the fact of “son preference”, which is a product of a patriarchal society, that itself is the underlying cause of sex-selective abortion, which in turn is one, although not the only, driver of a skewed sex ratio.

Deeply held cultural values and beliefs – even hateful ones such as misogyny – are slow to change and difficult, if not impossible, to legislate away. The correct response is not despair, nor a knee jerk appeal to barbaric punishments such as the death penalty or “chemical castration” of convicted rapists. The copious mass of evidence from the United States does not unequivocally support the widely held idea that capital punishment serves as a deterrent to future crime.

The unglamorous but right answer is to alter the incentives that would-be rapists and other assailants face when they contemplate a sexual assault, tilting them away rather than toward such crimes. The fact that, as of last year, only about a quarter of cases against alleged perpetrators of rape resulted in conviction is a telling statistic. What is the use of ratcheting up punishment from life imprisonment to the death penalty if there is still a three in four chance of an alleged perpetrator walking away scot-free? Whatever one's view on the morality of capital punishment, the salient point is that its efficacy as a deterrent is in serious doubt given such grim statistics.

While the research on the deterrent effect of severe punishment is hotly contested, as noted above, what is incontrovertible is that a criminal justice system which would efficiently and fairly mete out justice under our existing legal system will go a long way toward reducing the incidence of such horrific crimes.

We need to draw the correct lessons from the recent tragedy and not allow our understandable anger and revulsion to point us toward superficially appealing but misguided solutions.

Vivek Dehejia is an economics professor at Carleton University in Ottawa, Canada. He is co-author, with Rupa Subramanya, of “Indianomix: Making Sense of Modern India”, just published by Random House India.

2 comments:

Chatty Wren said...

You have raised some relevant points, am linking this to my blogpost.

Anonymous said...

Dear Sir,

In response to your post I had arrived at this conclusion some time back while working on my Pension Phd thesis using slightly different logic but never tried to verify.

Because I was closely looking at demographic data, and being from a family of doctors, I could understand that falling female-male ratio, particularly in Northern belts of India has resulted in what is know in medical science as Sexual Tension (ST). Aggressive behaviour due to ST is also observed in mammals and has been recorded by Zoologist. Falling female-male ratios decreases the probability of mating and hence ST. That could explains the negative correlation in your data. Also crime rate is highest among males in prime reproductive age 17-35 as observed in your data further supports my theory.

Hence, in the on going debate stronger laws will only be short term solution. Correcting female-male ratio by strongly clamping female feticide is a true long term solution in my opinion.

SH