Sunday, May 18, 2014

GE 2014: AAP's Delhi Debacle - How bad is it really?

One of the big disappointments in the General Elections 2014 was AAP's so-called debacle in Delhi. Having delivered a stellar performance in the recently concluded Assembly Polls (AE 2013), where it uprooted the 15 years old congress government and emerged as the second largest party within a few seats range of a majority - its inability to open account in GE 2014 (in Delhi) has left AAP supporters disappointed and from the news it seems AAP party workers bewildered and badly shaken (their demand from AAP leadership to again form Govt in Delhi without seeking a fresh mandate reeks of desperation and a sense of helplessness). So these developments left me wondering how bad is the drubbing really in Delhi for AAP? Would they had fared better had General Elections been held in Dec 2013 when supposedly AAP had a much better public perception? What really has been the impact of AAP's sudden decision to resign? Of course we all know that AAP's vote share has actually improved (though marginally from 29% to 33%) from last election but still a duck makes one wonder what went wrong? To me the funny thing is that when we talk about AAP's amazing Debut in Dec 2013, we all seem to forget that BJP was still the single largest party in Delhi - its almost as if BJP's performance has no memory share! Such was the impact of AAP's performance that perhaps it got stamped in our minds as a much bigger victory than it really was. So to understand this better, I decided to do a little number crunching - I took the AE 2013 numbers (by assembly constituencies - AC) and rolled it up by corresponding parliamentary constituencies (PC) to examine how AAP would have fared had general election been held in Dec? Would it have done better? Simply put the answer is yes.

If you look at AAP's vote share rolled up to PC level based on AE 2013 - AAP would have won 2 seats Chandni Chowk and New Delhi though BJP would have still come on tops in all the other seats. More importantly, when one looks closely at these numbers, one has to appreciate that BJP would have won with comfortable margins ranging from 4% to 11% while AAP would have won with a slender margin of 1% in Chandni Chowk and only in New Delhi it would have commanded a margin of 5%. Also, if you examine this New Delhi PC closely, this 5% margin is largely attributable to Arvind Kejriwal's stupendous 25.8K votes win over Sheila Dikshit and if you were to remove the New Delhi assembly constituency votes (assuming that AK would have been battling Modi in Varanasi) then the victory margin drops to 2%. Net net a victory by a whisker. So perhaps AAP's tally would have been better but nothing spectacular. 

Now comes the bigger question, did AAP's sudden exit in Delhi erode voter confidence and is the drubbing a result of that. A look at the vote share change data at PC level provides some insights into that.

We all know that AAP's vote share at an overall level has improved marginally in this election (from 29% to 33%) but interestingly, that improvement seems to be well spread out (rather than being concentrated on a single seat which then would have been attributed to that particular candidate), this points to that the fact that AAP is still widely perceived as a highly acceptable alternative. Importantly, the improvement in vote share ranges from 4% to 13% while the drop ranges from 1% to 3% except in New Delhi where it dropped by 8% (perhaps people punishing AK for abandoning the seat?). So my key take away is against the widely held perception that AAP has been punished by Delhi voter for a sudden exit, they seem to be very much there where they were in Dec'13 -perhaps this vote share would have been much better had AAP stayed in Govt? Who knows? But yes the damage doesn't seem to be as bad as we think it to be. As an aside, it would have been interesting to break this GE 2014 data by ACs to see how many seats AAP would have won had Assembly polls had been done simultaneously - however, that data is not available publicly, I am sure AAP and others must already be doing that maths (in fact, according to a news report BJP was on top in 60 out of 70 seats - so perhaps AAP would have been decimated!!).

So if voter didn't really punish AAP, then what went wrong? The answer is simple - Modi wave and the fact that most of the voters in Delhi had already decided that for Assembly it's Kejriwal and for Lok sabha it's goona be Modi. BTW AAP is not the only party which has been annihilated (actually annihilated is a wrong term in the context of AAP for there was nothing there to destroy but I am trying to be a lil dramtic!) by this unprecedented election results. Given that parties like BSP, left, SP, MNS, JD (U), RJDs which had their fixed strategic voter bases have lost so badly - I don't think AAP has done badly at all - they just couldn't make it to the podium. And the impact of this wave and voter mindset is evident when one looks at change in vote share for BJP (which has come mostly at the cost of Congress and not AAP btw)

 So to conclude, in my view -
  • AAP still needs to do a lot more work before it can stake a serious claim on a Delhi Lok Sabha seat - it definitely would not have been an easy win in Dec'13 either (except in case of AK)
  • Delhi has not punished AAP for a sudden exit - however, if the current trend holds AAP may still suffer in Delhi assembly polls more due to after effects of a Modi wave than anything else - so AAP needs to get its act together fast to retain Delhi.
  • A zero in Delhi was not AAP's doing but simply the effect of Modi wave, for a young party (which also has been the favorite whipping boy of other parties and media ) to still withstand such a massive vote shift is an achievement and you should be patting yourself on the back rather than being pensive about it
  • A drop in New Delhi vote share shows the importance of AK for a Delhi win - AK you need a home base what better than New Delhi as your constituency?

Tuesday, May 13, 2014

Indian Credit Card Market - A snap shot

While trawling through the RBI website for a totally different purpose I landed on some very interesting credit card data. Since I started my career in the cards industry - I could not resist running a quick analysis to understand how the industry is faring and got some interesting insights. Check it out!! (Disclaimer - Pls note that this is only based on one month of RBI data - Jan'14 so there might be gaps due to reporting issues etc.)

Cards in Force

As on 31st Jan 2014, there are 18.9 Mn cards in force with top 10 issuers contributing a mammoth 94% of the cards in force. While cards issued or in force by itself is not a good determinant of market share (which should be based on spend or active cards - unfortunately, data on active cards is not available with RBI though one can make some deductions based on the spend data) it definitely does highlight who are the aggressive issuers in the market (as to how they are performing is another issue which I have tried exploring in next section). And as expected the private Indian banks viz. HDFC and ICICI come on top closely followed by SBI - together these 3 banks account for ~58% of the total cards issued.

India's tally of 18.9 Mn cards pales when compared to China's ~331 Mn cards (reported as on end of 2012 Source: Bloomberg) and US' ~536 Mn (Source: Bloomberg) though some source also place US' cards in use to about 1.5 Bn (making it a whopping 3.5 cards per person). A look at RBI data shows that the number of cards issued has remained almost stagnant around 19 Mn mark - in fact it has fallen from about 19.6 Mn to current levels driven largely by HDFC's card base shrinkage between May to Aug last year (check out the story here -Business Standard's Story on HDFC's card portfolio shrinkage). Clearly Indian Bank's aggressive card sales followed by a heavy losses on that unsecured lending has dampened the spirits but this is also a pointer to the potential this market has to offer as it is clearly far far away from a saturation point. Just the debit card base of these issuers is 175 Mn (excluding Amex which is just a credit card company) and if we also exclude SBI then it stands at around 58 Mn - clearly some of them would be credit worthy!

Card Spend

For the month of Jan' 2014, the total spending on the card stood at Rs. 141.62 Bn (or assuming a USD:INR rate to be 60 ~USD 2.36 Bn). Given that this figure would include holiday spending, caution needs to be exercised before extrapolating this number for the entire year - for instance the spending in Dec'13 was around ~13.5K Cr while in Feb'14 it was 12.7K Cr. However, what we are more interested in is how this spending is split between different players. Once again this number might be slightly skewed due to seasonal offers - but would still give a good idea of how these players fare on driving spend.  A quick look at the charts reveal that while HDFC - the topmost issuer of cards also tops the spend volumes, the MNC banks Citi and Amex are 2nd and 3rd  even though they account for 13% and 3% of the cards issued respectively (once again since this is only based on Jan data - pls refrain from drawing a long term conclusion on this basis though interestingly the Feb data also shows a similar trend). ICICI the second biggest card issuer is at 4th position in terms of spend followed by SBI and Axis. They are then followed by MNC banks Stan Chart and HSBC. The numbers which stick out are ICICI and SBI who inspite of having 17% and 15% of the card issued have a spend share of 12% and 11% respectively and have been displaced by Citi and Amex with a spend share of 17% and 12% respectively. A part of this can be explained by inactive card base of these banks ie. ICICI and SBI which then also demonstrates that enough is not being done by them to get those cards activated - if you consider them worthy enough to have cards then you must make an effort to get them to spend on those cards - unfortunately that aspect can not be explored here due to lack of data. In addition, the premium customer base of Citi and Amex as compared to ICICI and SBI is another reason how this difference can be explained. However, while for SBI given that it is a aam junta bank, this difference is understandable, if I was a ICICI shareholder or part of the management team - there would have been some tough questions asked especially in light of HDFC's stellar performance - their card base should be pretty similar - in fact, given HDFC's bigger base one would expect the problem of inactivation and non-premium customers there to be higher but they seem to have successfully navigated the same. Similarly for MNC banks HSBC and Stanchart, while given their card base and spend, they don't seem to be doing too badly, a comparison with Citi and Amex - where again the customer base would be very similar points to a big room for improvement for these banks.

Now the spend share for any card player is driven by card activation and spend volume on the active card. While due to lack of data we cannot explore the former, we can definitely look at latter to understand a high level spend behavior ie. how much on avg their card members are spending which in turn is a function of  how many times do they transact in a month and how much do they spend on each transaction. These metrics should provide some insights on the card base of these banks as well as the effort they might be putting in incentivizing spend. The following chart on avg spend per card (pls note that this or any of the following metric does not exclude inactive cards) reiterates what we have seen so far where Amex with a small card base and high spend is at top (and way beyond the reach of others) at Rs. 26.8 K per card followed by Citi and HDFC at Rs. 9.8K and Rs. 8.1 K respectively. Not surprisingly, ICICI and SBI are the bottom two with an avg spend of ~Rs. 5.3K respectively (all data is for the month of Jan only). Perhaps for SBI and ICICI, this number may go up dramatically if we were to remove inactive cards (and may be they should consider a purge similar to that done by HDFC in the middle of last year - mentioned earlier in this blog prior to which HDFC's per card was also around Rs. 5.6K). Readers might be wondering why should banks be bothered about inactive cards as long as the credit card business is profitable and that perhaps this entire analysis should be done on active card base only to do a better comparison. While I cannot agree more with using active card base for this analysis (with proportion of inactive cards serving as another metric), I think inactive card base is an important reflection on the bank's ability to manage its portfolio. Even if the portfolio is profitable, inactive cards would be a drag on this profitability due to maintenance cost. More importantly, if these banks are not taking a stock of the situation and taking remedial measures (in how they market and underwrite), they might just keep adding more inactive cards to this portfolio which basically means that they are wasting their marketing and underwriting resources.

Transactions per card

As mentioned earlier, one of the key spend driver is how many times does a bank gets its customers to transact on the card. Banks have been working overtime to get their customers to move their cash and other forms of spending on cards. While traditionally credit cards have been used for entertainment and shopping, banks are now trying to capture payments such as bill payments and grocery spends - a recurring monthly spend with assured charge volume. To get such spends a number of incentives have been lined up such as cash back offers, discounts, loyalty rewards and of course, promotions (such as win a movie ticket if you spend on grocery or additional reward points if you transact X times in a month). Therefore, a comparison of transaction per card is a good measure of the effectiveness of these incentive programs especially since all of them would be competing for an increased share of the same wallet in most of the cases (all of us hold cards from more than one bank but would prefer to spend using those which give us better incentive) . And when one looks at the comparison - not surprisingly Citi comes on top followed by HDFC and Amex while other are in a similar bracket of 2.1 to 1.9 transaction per month (except Kotak) - the same trend has been observed in other months as well. Citi clearly seems to have got its act together with promotions like Citi discount on food and products like premier miles (where you can convert loyalty point to but any airline ticket) is clearly helping them drive transactions. While the transaction per card may be a bit higher for the month of Jan due to festive season - it still clearly sets a minimum threshold of 2.5 and an aspiration band of 2.5 to 3 transaction per card for other banks to strive for.

Spend per transaction

While transaction per card is largely a function of incentives one has to offer, spend per transaction is driven largely by the kind of customer base a bank has and then to an extent by the kind of incentives it has to offer. For instance, an option to convert big ticket purchases to EMI or promotions like win an ipad if you spend more than Rs. X thousand may prompt one to move a big ticket purchase to that card  (if being planned), but typically this would give a spurt only during certain months such as diwali etc. (when such promotions are floated and may increase the spend per transaction) whereas a round the year high spend per transaction is driven more by lifestyle choices which in turn is driven by the economic strata the customer comes from. This in turn means that banks need to pay attention to their marketing strategies by targeting the right customer segment and by developing the right products - but only if they prefer quality over quantity - something which will allow them to capture a lot bigger share of the credit card spend without having the need to have an expanded base as in the case of Amex. This may also help these banks optimize their marketing cost and cost of maintaining a much smaller base - though it may get offset by a premium service that they may have to offer (perhaps a topic for next blog?). Clearly, Amex's high spend share comes from it leading the charts with Rs. 9.6 K per transaction which is about 2.6 times the next best bank - IndusInd (a surprise entry) at Rs. 3.6 K per transaction. While it is not surprising that Amex comes on top having already seen the spend per card data, the difference between Amex and others is a bit surprising especially since it also comes in top 3 in terms of transaction per card. This piece of analysis also throws up some more surprises with Amex followed by IndusInd, Axis, Kotak and Stanc which can be taken as an indicator of the quality of their customer base while Citi and HDFC are 3rd and 4th from the bottom respectively. For players such as IndusInd, Axis, Kotak and Stanc this might throw up an interesting insight where besides focusing on expanding their card base (which they should anyways do given the penetration levels and hence the opportunity we saw earlier) they should maintain and further improve on the quality of the customer base. Interestingly for IndusInd, Kotak and Axis (the bottom 3 banks when it comes to transaction per cards), this also means that by focusing on improving transaction per card for their current base may also give them great dividends - they are already sitting on a gold mine and just need to start digging more. Coming back to Amex, what really intrigues me is the level of spend per transaction that Amex is able to extract especially since an Amex card holder would invariably be a customer of one of these banks and definitely be holding a credit card from one of these other top 10 players - what is it that Amex is able to offer which these banks can't to be able to drive such high spending? Another explanation could be Amex's corporate card program where a lot of big ticket travel spend might be coming on the cards but then other banks also have similar programs and in fact, given that they have other banking relationships with corporates as well (which Amex cannot have), they are in a much better position to leverage their relationship and capture this spend. Clearly something is amiss here either Amex's offering is truly unique or banks are just oblivious of this opportunity.


While this entire analysis has focused on spend, we have completely ignored the other big revenue driver for the credit card business namely the interest income from revolving credit. It is quite likely that banks showing up as laggards here viz. ICICI and SBI might come on top if we look at revolve volumes and income thereof. However, in my view, a credit card business should look at merchant discount coming from spend as its main revenue driver with revolve income only as icing on the cake - in fact, if my revolve incomes starts going up - I would be very worried about the quality of credit given the unsecured nature of lending. Once again my interpretation here are based on one month data so one needs to be careful before generalizing but some points do stick out -

  • With a card base of only 18.9 Mn - there is a lot of potential that this market still holds for issuers
  • HDFC, Citi and Amex seem to be ruling the market driven by primarily by large card base in case of HDFC, high transactions per card in case of Citi and high spend per trasnactions in case of Amex
  • ICICI and SBI seem to be laggards might be due to their large inactive card base and can take a cue from HDFC and can make their base leaner and more productive
  • The smaller Indian players in this segment within top 10 viz. IndusInd, Kotak and Axis seems to have a great customer base - just get them to take the card out more often!
  • Stanc and HSBC - you guys are stuck in middle - think on what can you do differently to play catch up with big brother Citi!