Description

“Credit seems to be associated with a particular kind of reflexive pattern that is known as boom and bust. The pattern is asymmetrical, the boom is drawn out and accelerates gradually, the bust is sudden and often catastrophic”

                      George Soros, Alchemy of Finance

At the latest Sohn Investment Conference David Stemerman of Conatus capital warned that the booming South African credit cycle is on the verge of a bust. He recommended shorting African Bank (ABL SJ) as a way to play this top-down trade.

 The trade has already made 23% since the conference,but, I think this is only the first leg down with more to come. David appears to have chosen ABL as a way to play his macro call. I agree whole heartedly with his macro call (I am probably even more bearish), but, I also think there are some specific African Bank issues which were not addressed which could make the short call on ABL even more compelling.

 In this write-up I plan to

1. Add color to the unravelling credit boom in South Africa.
2. Discuss issues specific to African Bank which highlights the precarious position it is in.
3. Review other short candidates which can be used to play the bust. South African consumer stocks are very expensive compared to their historic multiples and there are a lot of candidates to choose from. I will, however, focus mostly on Capitec (CPI SJ), ABL’s largest rival.

Unsecured Credit

Before I get to the body of this report I need to get the definition of “unsecured credit” out of the way

Unsecured credit is an instalment loan which required no collateral.  The debt is very expensive. Typically a customer borrows R10 000 and is required to repay at least R30 000 over three years.

As far as I am concerned customers who use these loans are either desperate or victims of slick salesmen who prey on the financially illiterate. Obviously instant gratification and keeping up with the Joneses does play a role, but, the amount of borrowing just to make ends meet should not be under estimated.

I thought the most efficient way of covering the macro ground is to give my opinion on the points made by Conatus Capital.

1.  The unsecured credit market has tripled in the last three years.

 True, South African banks have thrown a party in unsecured lending that would make Angelo Mozilo blush.

The table below shows how the unsecured credit book has boomed from ZAR55bil in 2009 to ZAR159bil in 2012. Looked at differently, unsecured credit is currently 11.1% of the total credit market compared to 4.8% in 2009.

2. Unsound lending practices

This is difficult to prove as no lender is willing to admit that their underwriting black box is not working. I think we all know the drill. Management tend to boast about the sophisticated dials the PhD’s have come up with to navigate the risks in the credit market. Then, some obvious fatal flaw a 5 year old could figure out pops up and blows the black box sky high.

 I am going to try and be that 5 year old and list some of the things that make you go Hhhmmm when it comes to underwriting standards.

• Loan growth has been exponential. ABL sold ZAR25bil of loans over the last 12 months, double the ZAR12bil sold in the 12 months ended March 2010. The main competitor, Capitec also sold ZAR25bil over last 12 months three times more than the ZAR8.6bil sold in the 12 months ended March 2010. Given that employment growth has been tepid over this period I am sure a reasonable 5 years old would go Hhhmmm.
• I know that commission driven agents are required to sell at least R1mil of loans per month. I can just imagine the tricks the incentive driven agents get up to to squeeze the loan through the underwriters black box. 
• ABL’s latest results show that the bad debt from their lower risk customers are tracking well above expectations. These customers are lower risk mostly because they earn higher salaries and have been aggressively targeted by all credit providers during the boom. Seems to me that the model may have missed that the low risk customers were targeted by every man and his dog. 
• Over the last few months ABL have increased the size of the collections call centre from 800 to 1300 collectors. This is a strong signal that something must have gone wrong in the underwriting department.
• The ex FD of ABL, Dave Woolam, recently warned that the “pool of debt we are seeing now will probably grow — possibly to above a 50% default level."

 http://www.bdlive.co.za/businesstimes/2013/05/12/from-boom-to-bust-warning-for-the-banking-sector

3. Consumers are spending 48% of their income on debt service.

 Not true, I think it is much higher than this. Discussions I have had with organisations which help people struggling with debt indicate that the ratio is closer to 70%.

 This is obviously a biased sample so let’s corroborate this with a little DD. On page 8 of ABL’s Sept 2012 results presentation there is a graph showing that loan applicants rejected by ABL had a average debt service cost to income ratio of 70%.

 The same graph shows that applicants who were accepted by ABL had an average debt servicing costs to income of 40%. But, and this is important, the 40% ratio is calculated before taking the debt servicing costs of the new loan into account. By the time the costs of the new loan are added in it is not difficult to imagine that the ratio goes well above 70% leaving very little for everyday living expenses.

 http://africanbank.investoreports.com/downloads/resultsreportspresentations/2012/ABIL-2012-PPT-and-fins.pdf

That there are people who have a debt service cost to income of 70% applying for credit clearly indicates that there are a lot of desperate South African’s using debt to put food on the table or borrowing from Peter to pay Paul. To date easy credit has supported this behavior, but, the credit cycle is turning and there are signs that the easy money is drying up fast. If this continues the true bad debt levels will emerge to shock lenders and investors.

4. There has been widespread fraud and banks have been complicit.

 True. The fraud has occurred on two levels

1) Affordability tests and

2) Collections

Affordability Tests: The National Credit Act (NCA) states that a loan can only be made after conducting an affordability test on the customer. In other words “thou shalt not make a NINJA loan”.

 No problem, all that needs to be done is to make a couple of adjustments on the application form overstating income and understating expenses and magic the loan is affordable.

 “In one case, a forklift driver, whose net monthly salary was R2201, was assessed as having R2100 to spend on debt, as the shop assistant had entered R99 as his "minimum living expenses". Lewis agreed to sell him goods worth R3109, which worked out to R8976 over 24 months, at R374/month.”

 “In another case, a Lewis credit agreement states that a kitchen hand who earned R2450/month at a coffee shop in Cape Town only needed R1 for "minimum living expenses", so "money available to pay debt" was R2449. She bought a five-piece dinner set (and R139 plug) for R190/month.”

 Source: http://www.fm.co.za/fm/2008/09/12/furniture-retail-fiasco

 Stories like these are frequently in the press, but, more importantly occurring in front of the dozy regulators nose.

The regulator has done very little since the new Act was implemented in 2008. Then in August 2012, following the Marikana tragedy (34 miners killed by police), things changed. The regulator stirred and became very excited about the widespread abuse in the industry.

 http://www.moneyweb.co.za/moneyweb-special-investigations/credit-regulator-swoops-on-marikana-lenders?sn=2009+Detail

Politicians who previously did not care a hoot about their constituent’s plight are suddenly all on the bandwagon making political hay and blaming the Marikana tragedy on the micro lenders. They have also warned that they are going to amend the credit act to tighten up on the affordabilty tests. This will probably include a cap on debt service costs to income.

The micro lending business model works as follows. Make the loan and then try and collect all the instalments. The most efficient way to collect is to establish when a customer gets paid and then to run a debit order against his bank account on the same day.

 If the debit order fails, then the account is directed to a call centre who start off with soft collection strategies which becomes increasingly unpleasant the longer the amount remains outstanding.

 If the call centre fails to collect the outstanding amount the account is handed over to external law firms for collection. These lawyers generally go to court to obtain a garnishee order. The garnishee order is an order from the court which instructs the employer of the customer to deduct the amount owing from the employee’s salary.

 Sounds great. But, what happens if the lawyer is not a model citizen and is prepared to game a corrupt court system or just fake his own court documents.

 ENS, a top 3 law firm in South Africa, decided to investigate the mess and found that 59% of garnishees were irregular. (Experts estimate there are 3million garnishee orders in the system. An incredible number if one considers that there are only 11mil people employed in South Africa.)

 http://www.moneyweb.co.za/moneyweb-special-investigations/sa-infested-with-debt-abuse  

Further, it has been hard to avoid headlines in the press with the same sad story “Platinum miner repays R11 690 on R1 000 loan.”  

http://www.moneyweb.co.za/moneyweb-financial/platinum-miner-repays-r11690-on-r1000-loan

http://www.moneyweb.co.za/moneyweb-financial/industry-scrambles-to-save-garnishees

Specific African Bank Issues

African Bank consists of two business units, banking and furniture. The units are closely related as the furniture operation essentially sells furniture on credit in order to originate loans for the bank.

As the following timeline shows the stock is already down 23% since the Sohn conference.

May 2, 2013: Profit Warning, stock falls from R29.00 to R23.90

May 8, 2013: Sohn Investment Conference, stock unchanged at R23.00

May 20, 2013: Interim results announced, market surprised by dividend cut, stock price falls from R21.06 to 17.20

May 20, 2013: Current share price R17.60

To date the ratings agencies have been positive on African Bank. This has allowed ABL to finance their frantic growth in the book by frequently issuing bonds to generous bond funds scattered across the globe. Obviously, this access could change over night if the ratings agencies change their minds. So yes I agree with Conatus the wholesale funding model is a risk, but, I think this risk is more severe than presented because of an aggressive accounting policy which never seems to receive much attention.

The accounting trick which overstates the equity of the bank works as follows.

ABL's loan book is disclosed as follows on the balance sheet.

Provision Doubtful Debt                        R10.4bil

Net Book                                            R48.4bil

Value of Written Off Book                    R  2.1bil

Net Book                                            R50.5bil

But, what does “Value of Written Off Book” mean in English? This is bad debt which was previously been written-off, but, because amounts are recovered every now and then the accounting geniuses want to take a portion of the witten-off book and put it back on the balance sheet as an asset. The R2.1bil represents this entry and values the written-off book at 14c on the Rand.

 Never mind that the accounting is not exactly prudent, 14c is way to expensive as everybody knows that written-off books typically trade at 5c in the Rand.

 Lets back this up

•          Capitec, the big rival across the street, value their written-off book at 4c on the Rand.
•          ABL's bad debt recovered over the last 6months totalled R109mil. If we annualise this we get R218mil per annum. Assuming that R218mil is recovered every year for 10 years then the DCF of the cash flow using a 10% discount rate is only R1.3bil. Much lower than the R2.1bil.

 I think the accounting is aggressive, but, more importantly I think the 14c is nuts. If the rating agencies/bond funds/auditors wake up to this and decide to plug in 5c the following adjustments will have to be made to the banking units balance sheet. (I am limiting this analysis to the bank’s balance sheet as this is the focus of the ratings agencies and bond funds.)

 The value of the write-off book falls from R2 100mil to R750mil. This will knock the value of the banking unit's equity 12% from R12 300mil to R10 950mil.

 These types of adjustments to equity can really spook bond markets. It is not so much the size of the write-down, but, the signal that is sent especially regarding management credibility. After all, banking is a game about confidence full stop and overstating assets/equity does not go down well. Obviously if the bond market shuts down equity investors will be kept awake at night fearing further dividend cuts and dilutive rights issues.

 I don’t want to get personal, but, the guy running the furniture division makes Ron Johnsons attempt at turning JCP look like an amazing success. How he has kept his job is a mystery. The CEO/Board seems to be blind to the endless parade of false promises.

 But, I am an outsider so what do my thoughts on the management team really count. A far more powerful signal comes from one of the founding members who started the company with the current CEO. It appears that he has sold all his shares. It has always been my understanding that these guys were joined at the hip and from what I hear the sale was driven by a loss of confidence in the current management team. Unfortunenately I cannot corroborate the above, but, I do think my sources are very reliable.

Valuation

• The written-off book is overstated by R1.3bil

• There is R5.5bil of goodwill associated with the struggling furniture division which I think should be written-off.

Following these adjustments the equity falls from R15bil to R8.2bil which equates to R10 a share. The current share price is R17.20 which means there is still lots of downside left for the shorts.

Other Short Candidates

The boom in credit has boosted the sales of South African retailers. Foreign investors have flooded into these stocks (most are at least 60% owned by foreigners) pushing PE’s a couple of standard deviations above their historic averages.

 The poster child of this irrational exuberance is Massmart (MSM SJ) in which Wal-Mart bought a 50% stake in 2010. The stock trades on a PE of 29 even though earnings have largely been flat since 2008. Further, they have just warned the market that operating profit is expected to decline over the next year. Before 2008 the average PE of the stock was 15.

 I have a lot more to say about the very  expensive retail sector, but, this write-up is getting too long. Please feel free to ask more in the message thread. For now I want to end off with a brief discussion of ABL’s biggest rival Capitec.

Share Price = R197

Market Cap = R22.7bil (ABL only R13.9bil)

Equity = R8.3bil

PTB = 2.7

Gross Book Value = R30.7bil 

Provision DD = 2.7bil

Net Book = 28.0bil

%52High = 89% (ABL = 44%)

To cut to the chase the biggest reason for the short is that they have grown their book at a world record pace and I just can't get the image of the bank throwing money at the customers  out of my head. The table below shows how loans sold ballooned from R6.2bil in 2009 to R25bil in 2013. It also shows that ABL have pulled back on sales in 2013, while Capitec have kept their foot flat seemingly unconcerned that they may be racing towards a brick wall.

Other reasons for recommending a short on Capitec include

1)      Stock is expensive trading at 2.4 times book.

2)      Stock is still close to it’s all time high, especially if one adds back recent dividends to the current share price.

3)      African Bank have 1300 people in their call centre, Capitec have less than 200 as they outsource most of their collections to law firms who collect aggressively. As I mentioned above the regulator is tired of the abuse of the garnishee orders and is threatening to abolish the system. This could be a serious blow to Capitec's collection platform.

4)      In 2005 Capitec was more of a payday lender making small one month and three month loans targeted at low end customers who could not make it to the next pay day. In 2013, 80% of the loans sold had terms greater than 12 months and 45% greater than 60months. This is a dramatic shift in the business model. Also, the longer term loans are much larger and have been targeted at a higher end customer who tends to earn a higher salary. As mentioned above this customer group was the low hanging fruit which everybody targeted in the boom. Not surprisingly ABL's latest results show that the bad debt experience from these customers is tracking well above expectations and I think it is just a matter of time before Capitec have to face the music..

5)      7% of Capitec’s loan book is exposed to the mining sector. The next few months are know as the “strike season” in South Africa where miners traditionally bargain/strike for increased wages. This year’s “strike season” is expected to be brutal as an upstart labour union (AMCU) which has quickly become the majority union on many mines as disaffected NUM members deserted this once dominant union for AMCU in droves. NUM is closely affiliated to the ANC which currently governs South Africa. AMCU seems to be a law unto itself so get ready for unions scrapping with unions, management scrapping with unions and all this under the auspices of a biased government who face an election next year. The first salvo has already been fired. NUM want a 60% wage increase from the loss making platinum mines. 

Conclusion

 Back to George Soros and the Alchemy of Finance.

 “The act of lending usually stimulates economic activity. It enables the borrower to consume more than he would otherwise….By the same token, debt service has a depressing impact. Resources that would otherwise be devoted to consumption are withdrawn. As the total amount of debt outstanding accumulates, the portion which has to be utilized for debt service increases. It is only net new lending which stimulates, and total new lending has to keep rising in order to keep net new lending stable.”

 The steep growth in unsecured loans sold over the last four years is shown below

            Dec 2010: R52bil

            Dec 2011: R83bil

            Dec 2012: R102bil

Remember these loans are expensive and it is not difficult to imagine that just from the R102bil loans sold in 2012 customers will need to repay over R300bil over the next three years, or R100bil per annum. With more and more credit taps being turned off and a jumpy regulator I expect the loans sold in 2013 to be signicantly below the 2012 level. This means that a large portion of the R100bil repayment due in 2013 will have to come from consumption or bad debt.

This depressing effect will erode collateral values (including salaries) making it even more difficult to service the debt. Before you know it you have a vicious negative cycle which brings you face-to-face with the possibility of a “sudden bust which is often catastrophic.”

Catalyst