In Between the Buyer and the Seller, Mint columnist Karthik Shashidhar uses concepts from financial markets to explore everything from medieval bazaars to modern platforms. The following is an excerpt.
A person doesn’t know true hurt and suffering until they’ve felt the pain of falling in love with someone whose affections lie elsewhere. — Rose Gordon in Her Imperfect Groom
Where all women are bots
If you are a woman in India looking to find a date through an online dating platform, you might be overwhelmed. You will find yourself constantly receiving messages from men you are highly unlikely to be interested in, and in the process miss out on messages from men who might actually interest you.
If you are proactive and try to find a date by going through people’s profiles, you might find that you have to go through numerous profiles to find a handful that interest you. The problem is not that there are not enough date-worthy men on these platforms—this issue is finding them and getting through to them. In order to find a man that a woman might consider worthy of a date, she has to spend a considerable amount of time and effort weeding through a large number of profiles who may not interest her at all. This makes the process of finding a date on the platform costly, and might put off the woman from using the platform itself.
On top of all this, a large number of men on Indian dating platforms have come to the “conclusion” that most women on these platforms are “bots”. This conclusion is borne out of the frustration of expressing interest in and sending messages to a large number of women, and yet being unable to find a single date. Some men, in their desperation to get a date, even use “extension apps” that enable them to simultaneously express interest in large numbers of women.
In practice, that hasn’t helped improve their chances of finding a date, and has instead fed into conspiracy theories such as women on the platform being “bots”. Another popular conspiracy theory is that dating apps have a “no matches” bug, which prevents users from finding matches. There is also the belief that one needs to periodically refresh the apps and their connection to Facebook in order get matches.
To be fair, men on dating platforms in India don’t have it easy either. With women being flooded with a large number of profiles and messages, men need to put in special efforts in order to stand out and attract their attention. Every aspect of men’s activity on these platforms, from the pictures they use to the tags with which they describe themselves, and the intro messages, needs special care. Apart from trying to be attractive, men on dating platforms have the task of conveying their credibility as well.
In other words, dating platforms in India have a significant congestion problem. Women have to wade through a large number of profiles before they can find men who they might want to date. Men, on the other hand, need to stand out in the crowd and ensure they get noticed by women, apart from whatever else they have to convey. Thus, both sides of the market face significant transaction costs, resulting in low market liquidity.
Online dating platforms are a relatively recent phenomenon in India compared to the rest of the world. In fact, dating is mostly an urban phenomenon as a majority of Indians still continue to find their long-term gene propagating partners through what is known as “arranged marriage”—where families take an active interest in finding someone their potential spouse. In extremely conservative communities, the people getting married don’t even have an opinion on the matter and are forced to marry the person of their families’ choice. There is also a gender imbalance here—women are more likely than men to marry against their wishes.
While the arranged marriage “market” (some people find the use of the word “market” for relationships repugnant, hence the quotes) has its own issues in terms of search and transaction costs, one fallout is its impact in the dating platform market. With women being less likely to go against their families and find dates for themselves, there is a massive gender imbalance in the user base of dating platforms. Within each socio-economic group, a larger proportion of men is likely to be on these platforms compared to women, because of which the dating market is loaded in favour of women. The second order effect, of course, is that women on these platforms find themselves flooded with expressions of interest, significantly diminishing the quality of their experience.
Romantic novels and movies of a certain type send out the message that a hero gets the heroine because he loves her more than any other suitor. Such stories wreak havoc with many a teenager’s life, sending them on an endless pursuit of a certain somebody, believing that the amount they love the somebody in question will enable them to get the somebody to love them back.
Such pursuits sometimes succeed, for love can be persuasive, and expression of love from one person to another can have a positive impact on the quantum of love in the opposite direction. In most cases though, these pursuits end up following the storyline of another type of books or movies—where the heroine spurns the hero in favour of a “third guy” despite the hero “loving her more”.
The problem, if you want to describe as such, with dating or getting into a relationship is that it is necessarily a “two-way deal”. Not only should you approve of your date or partner, but the same potential date/partner also needs to approve of you in order for you to have a deal. Literature and poetry might glorify unrequited love, which might have produced such literature in the first place, but it is of little significance as long as real life is concerned.
As economist Al Roth puts it in his book Who Gets What And Why, the question of who gets what in a commodity market is determined by price alone. So a painting at an art auction goes to the highest bidder—the painting itself has no say in the matter. Similarly, in stock markets, traders who offer the most attractive prices have the best chance of winning the trades—counterparties seldom discriminate based on who they are trading with. Retail stores nowadays typically have fixed prices, and shopkeepers are willing to sell goods to anyone who is willing to pay the stated price.
However, not all markets function this way. A government will not necessarily offer a contract to build public infrastructure to the lowest bidder—it should be convinced that the bidder is likely to finish the construction in time, and with the requisite quality. Similarly, banks give loans not necessarily to borrowers who offer to pay the highest interest rates, but to those with whom they have a relationship, based on which they can be highly confident of the loan being repaid. Colleges don’t necessarily offer admission to students willing to pay the most fees, and employees don’t necessarily take up the highest paying job offer.
Roth describes such markets, where factors other than price have an important bearing on who gets what, as “matching markets”.
Going by the story arc of romantic novels and movies described above, one might be tempted to think that the market for relationships is a commodity market, with the currency of exchange being the “quantum of love”. If this were strictly true, a person would choose as partner whoever “loved her most”, without any regard for how much she loves this potential partner in return.
In most cases, though, this is not the case. While the amount of love received from potential partners matters, what matters more is the love people feel for their partners. And if both partners feel that way (weigh their own love for the other more than the love the other feels for them), we can assume that a date or a relationship or a marriage happens if and only if both parties approve of each other to a certain extent.
This is not a book about the philosophy of relationships. This is a book on markets. All the above “build-up” goes to illustrate that the market for relationships is a “matching market”, where the striking of a deal is dependent on both parties approving each other, with a further complication being introduced by the fact that the entities being “traded” are the transacting parties themselves. And this has important consequences as far as the design of relationship platforms is concerned.
The fact that both transacting parties need to approve of each other means there are three reasons you might fail to get a date—either you may not approve of anyone, or nobody approves of you, or that people you approve of don’t approve of you (the same applies for other relationship stages as well, with the intensity of approval being different). Hence, from a platform’s point of view, it is not just necessary to bring men and women together—it also needs to be designed in a way that people who might mutually approve of each other can find each other.
Dating platforms vary by how they seek to improve matchings, but there has been no magic formula yet. Tinder takes the buzzword “hyperlocal” to a new level by encouraging people to find matches in a small geographical radius. This can be useful in situations such as bars or parties where people might be secretly checking each other out but aren’t sure if they want to approach. This method also relies on the assumption that people who end up at the same party or bar are more likely to approve of one another than a random pair of people in different locations.
Platforms such as OKCupid and TrulyMadly, on the other hand, use questionnaires and tags to make recommendations. Users are asked to describe themselves and their ideal partners using a series of questions or tags (exact implementation varies by platform), which the platform then uses in order to make matches more efficiently.
India-focussed TrulyMadly also makes assumptions on a user’s choices based on their demographic information. Once a user has signed on to the application, the platform uses the demographic details available to quickly construct defaults on how the partner should be. The site assumes, for example, a particular range for the age difference between partners. The site also assumes that women like to date men who are as much, or more, qualified as they are in terms of education.
While some of these measures might seem overbearing and paternalistic for a neutral liberal observer, what these platforms are doing based on these matches is to make use of data from matches already made to learn about their future users’ behaviour, and use it to make better matches and cut congestion.
The high, and growing, user base of apps such as OKCupid, Tinder and TrulyMadly indicates that some of these measures by the platforms to improve liquidity are actually bearing fruit. Yet, men on these platforms continue to be frustrated in their attempts to attract women’s attention, and women continue to be overwhelmed by the sheer volume of responses.
In order to understand how liquidity on these platforms can be further improved, it is instructive to look at how the market for arranged marriages in India has evolved through the last two centuries.
Hall’s Marriage Theorem
Imagine you are a high school teacher who has taken upon herself the task of pairing up students for the school prom. There is an equal number of boys and girls in the class, so such a pairing should be theoretically possible. The only issue, though, is that high school students can be incredibly fussy, and have strong preferences on who they want to be paired with.
So each boy is willing to take as a partner only a subset of the girls in class. Likewise, each girl will take as a partner only a subset of the boys. Each boy and girl has his or her own set of preferences, so the set of girls liked by one boy may not match the set of girls liked by another. A pairing is satisfactory if and only if each boy and girl approves of his or her partner—even a single boy or girl disapproving of his/her partner can render the entire matching process moot. The question is if a satisfactory pairing is possible at all.
The answer to this problem is a seminal result in Graph Theory (a branch of mathematics) which is known as Hall’s marriage theorem. Proposed by Cambridge mathematician Philip Hall in 1935, the theorem states that an efficient pairing exists if and only if for every subset of boys, the number of girls they collectively like is at least as much as the number of boys in the subset, and for every subset of girls in class, the number of boys they collectively like is as many as the size of the subset (a group of boys collectively like a girl if at least one of the boys likes the girl, and similarly for girls).
That Hall’s condition is necessary is not hard to establish. If a set of three boys together like only two girls, for example, it is easy to see that at least one of the boys has to go without a match. Hall’s genius lay in the proof of sufficiency—to show that if the condition is satisfied for every subset of boys or girls, a stable matching indeed exists.
Hall’s theorem is relevant to our discussion on relationships and dating for two reasons. Firstly, the fact that it has come to be known as a “marriage theorem” shows the importance of romantic relationships in the study of matchings in Computer Science and Economics. Secondly, it helps show how users’ fussiness can have a massive bearing on the ease of finding matches.
Assume, for example, that your school has two sections, with the behaviour of students varying significantly by section. At one extreme, Section A is so tightly knit that each boy in that section is willing to be partnered with each girl of the section (and the other way round). Pairing up students of Section A for the prom becomes a rather trivial exercise, since any pairing will be satisfactory.
Section B, on the other hand, has evolved differently, and let us say that each boy in the section likes exactly one girl, and each girl likes exactly one boy. While this is rather extreme even by high school standards, it is not hard to see that pairing up this class for the prom is contingent upon each boy-girl like being mutual. Even if two boys like the same girl, or one girl doesn’t like the boy who likes her, no pairing is possible.
While the real ease of matching lies somewhere in the spectrum between these two extremes, what Hall’s Marriage theorem shows is that the more fussy people are the harder they are to be matched. While it is not common to come across the exact situation of Hall’s Marriage Theorem in real life (since matches happen asynchronously), it helps us better analyse the ease of finding dates.
Wanted: Brides and Grooms
Long long ago when most people lived in villages, people couldn’t be fussy about who they dated or fell in love with or married. The inherent structure of the village society meant that there were few people of the opposite sex, and a comparable age, that people came across during the course of their lives. Moreover, rural societies had rigid caste or class constraints, which further limited social interaction and the pool of marriageable counterparties.
While events such as fairs and festivals brought together people from a larger geographical area giving them an opportunity to mingle, most people simply ended up marrying people chosen for them by their families. These matches would typically be brokered by influential citizens of the area such as priests and village headmen, and the people getting married seldom had a choice.
Urbanisation increased the pool of people one could marry, but the information problem remained. How could you recognise and locate people of the opposite sex and a comparable age, whom you could potentially date or marry? The relative anonymity of urban life meant that traditional village-like networks involving priests and headmen were no longer available for matchmaking. Differential migration patterns (men were more likely than women to relocate to cities) also led to a market mismatch, and increased difficulty in getting matched.
It was time for new intermediaries to step in to help make matches in urban areas. Leading the way was the humble newspaper, in the form of classified advertisements.
It took less than ten years after the invention of the modern newspaper in the 1690s for people to start placing classified advertisements seeking brides and grooms. These advertisements were initially unpopular, and some people thought respectable people would not advertise in a newspaper for a bride or groom. Sheer economics and efficiency, however, has meant that newspaper advertisements have remained a popular tool to find partners to this day.
In India, too, the newspaper was the platform of choice for finding long-term partners for much of the 20th century (classified pages of a newspaper act as a clearinghouse to connect people who place advertisements to those who read them). As they got more popular, however, there was a problem—congestion. With each newspaper edition having a large number of advertisements, readers had to go through all of them in order to find people they wanted to match with. The large number of advertisements also meant that people posting advertisements now had less of a chance to be discovered by readers.
And then classified advertisements lived up to their first name. With most Indians preferring to get married to people of their own caste (a survey found that over 80% of Indians marry within their caste), newspapers simply organised advertisements by caste, thus cutting down the time taken to search listings.
While newspaper advertisements remained popular through the 20th century, and they were helpfully organised by caste and community, the search cost remained significant. Moreover, given the number of newspapers in circulation, and that these advertisements were published only on one particular day, the platform was quite fragmented.
A reader of a newspaper could only access listings placed in that particular newspaper on that particular day. And a listing would only be seen by a subset of people who were reading the newspaper on that particular day. Searching classified advertisements was also inefficient.
Thus, it didn’t take long after the coming of the internet for online matchmaking platforms to crop up. Two prominent Indian platforms, Shaadi and BharatMatrimony.com set up operations as early as 1997, when only a handful of Indians had access to the internet. These platforms mainly targeted the rather substantial Non-Resident Indian (Indians living abroad) crowd.
Online matchmaking platforms scored over newspapers in multiple ways. Not only was there less fragmentation, but advertisements would also be visible for a longer period of time, improving the probability of matching. Moreover, online advertisements were agnostic to location, and potential brides and grooms in different cities could connect to each other using such platforms, boosting liquidity.
In the early days, these online matrimonial platforms also doubled up as dating platforms. Over time, however, the traditional arranged marriage culture of India had meant that these platforms had been taken over by “parents”. It became increasingly common for people listed on the platform to have their accounts maintained by their parents or other elder relatives, making it hard to locate counterparties who represented themselves. Soon, it became next to impossible to find conventional dates on these platforms.
In parallel to classified sections of newspapers, the 20th century in India saw the growth of yet another intermediary in the matrimonial market—the so-called “matrimonial exchange”. Some of these exchanges were run by charitable organisations, such as temples and religious trusts, while others were more commercial, charging participants for being listed. The way they differed most from newspaper classifieds was in their symmetric design.
A prospective bride or groom could gain “membership” in such an exchange upon payment of a fee and submission of personal details in a prescribed format. Membership conferred two benefits onto members—on one hand they would get listed on the exchange while on the other they would get to inspect profiles of potential counterparties. This membership driven model meant that exchanges offered more privacy compared to open classifieds in newspapers or online exchanges.
A long time ago when I was single and considering marriage, a relative had listed me on one such exchange. Having paid the listing fees, filled up the form and got myself listed, I now had the privilege to inspect counterparties on the exchange. I remember going there one Saturday morning, having been told by the concerned relative that there were “lots of women” in the exchange database.
I was not to be disappointed. Soon after I went to the exchange, I was handed four thick binders, which were full of profiles of women who belonged to my caste and who wanted to get married (organising binders by caste is a way to cut congestion). Each page in the binder had the profile of one woman, including a full length photo and bio data (I’m sure my profile had ended in a similar binder, full of men). The standard format meant that it was easy to parse.
Looking back, the experience was not very different from that of a modern dating platform, where you swipe left and right as potential counterparties arrive on your smartphone screen. Only in the case of the matrimonial exchange, instead of a smartphone, you had binders full of women!
Between the Buyer and the Seller is available in both print and Kindle formats on Amazon.
Comments are welcome at email@example.com