Wednesday 25 January 2017

India: Moody’s and ICRA See “Subdued” Prospects for India’s Banks

Sometimes Even When You See Something Clearly, You Think It Wise to be Indirect

Just when I was recovering from The National Bank of Ukraine’s festival of euphemisms about PrivatBank, along come Moody’s and its Indian affiliate ICRA to once again remind AA that his attempts are easily upstaged. 

In a report released on 9 January, Moody’s and ICRA summarize their conclusions about the country’s banking sector with the phrase “see subdued prospects for India's banks“.
Why is AA “skeptical” and inclined to a stronger term than “subdued”?  Perhaps “dismal”?

Three factors.
First, Indian banks—particularly public sector banks or PSBs—have a reputation for under-reporting NPAs.    Favorite techniques were refusal to recognize NPAs, disguising bad loans via restructuring and/or making new loans to pay interest on past due loans.   Former RBI Governor Raghuram Rajan launched a “crackdown” in 2015 to curb under-reporting of NPAs. 
Performance suffered.  The decline was chiefly due to increased provisioning in 2016 and the related impact on net interest margin.   According to RBI’s Report on Trends and Progress of Banking in India,  Operations and Performance of Scheduled Commercial Banks Table 2.1, banking sector return on assets for 2015/2016  was 31 bp and ROE 3.59% compared to 2014/2015’s ROA of 81 bp and ROE of 10.42%.   Public Sector Banks—some 70% of banking assets--fared even worse with negative ROA and ROE in 2015/2016.  
Second, Indian banks have also traditionally under-reserved their declared NPAs with provisions averaging roughly 40% of total NPAs.   According to RBI Handbook of Statistics of the Indian Economy Table 65, 2015 reserving levels were at 46%.   Unreserved NPAs were some 20% of 2015 capital (Table 64). 
It’s hard to tell what happened in 2016.  Much higher provisions were taken, but more loans were recognized as NPAs and restructured loans are now to be included in that figure.  What’s the net effect?   
Sadly, RBI data on NPAs is available with a roughly 12 month lag.   See Table 65 in the RBI’s “Handbook of Statistics”.  Latest figures are from September 2016.  Other RBI reporting has detailed bank-by-bank analysis but the latest data appears to be March 2016. 
Without RBI statistics on both NPAs and provisions, it’s not possible to determine if the provision coverage has increased because both NPAs and provisions have increased.  
Third, low provisioning levels are particularly important because NPA recovery is traditionally very low in India.  According to RBI’s Report on Trends and Progress of Banking in India,  Operations and Performance of Scheduled Commercial Banks, Table 2.2,  in 2016 Indian banks recovered roughly 10% of NPAs versus 12% for the previous year.  
What this means is that recoveries are unlikely to make up provisioning shortfalls to any meaningful extent.   Provisions then are more critical than in jurisdictions where average recoveries are in the 40 to 50% range. 
It’s hard for AA to imagine that during 2016 Indian banks cured decades of bad practice and bad underwriting.  Trump Tower like Rome wasn’t built in a day, though it is by some Twitter accounts better.  And banking sector cleanups generally take more than a single year. 
Moody’s/ ICRA seem to agree. In their press release, they project single digit ROE for 2017 and 2018 and note large capital needs particularly among PSBs. 
A case of JPMorgan “Jakarta” fever?  Or euphemism?  
And finally a tip of AA’s enormous tarbush to ICRA SVP Karthik Srinivasan for combining “dent” with “profitability matrices”.  See link to Moody’s / ICRA press release. Shabash!

Sunday 22 January 2017

Golden Curtains

New Administration, New Curtains in the Oval Office
Contrary to some media reports, these do not appear to be shower curtains.

Friday 20 January 2017

KHCB: Credit Metrics Part 3: Collateral Coverage


Post Foreclosure Sales Prices May Be Less than Appraisal Values

We ended the last post on KHCB’s credit metrics with some theological speculation.

Since AA’s province is finance and not theology, let’s look at collateral coverage.  Maybe KHCB is so collateral “rich” that classifying some loans past due 180 days as unimpaired makes perfect (credit) sense.

As outlined above, KHCB’s collateral position is not sufficiently “robust” to compensate for other weaknesses in its lending portfolio discussed in earlier posts.
  1. Collateral is concentrated in real estate. That poses at least two problems. Illiquid assets like real estate are difficult to sell quickly and/or at full price, exposing lenders to loss.  Real estate is interest-sensitive. With real estate 85% of collateral, KHCB is particularly vulnerable to increases in interest rates. 
  2. Compounding this problem, KHCB’s real estate collateral is primarily located in a single, very small market—Bahrain—, magnifying the inherent risks of taking illiquid assets as collateral. 
  3. KHCB also seems to be relying on real estate to collateralize non-real estate loans thus increasing the bank’s overall exposure to real estate.  This appears to contradict GFH’s stated goal of reducing “land-based” exposure.   
Now to the detail.  

Key areas for investigation: 
  1. Percent of portfolio collateralized and trends in collateralization.
  2. Types of collateral.  
General introductory notes:
  1. Note 34 page 72 (IFRS) and Note 4.10 page 93 (Basel Pillar III) are the sources for this post.
  2. I have relied primarily on the Pillar III note as the IFRS note really doesn’t provide the same level of detail.
  3. However, both have an apparently erroneous and misleading statement regarding collateral coverage of the portfolio.  This error appears only as a number in the Pillar III Note 4.10.  Note 34 spells out the error:  The average collateral coverage ratio on secured facilities is 107.80% at 31 December 2015 (31 December 2014: 109.49%).”  
  4. As Pillar III Note 4.10 shows this ratio was determined by taking the value of all collateral and dividing it by outstanding exposure. 
  5. Two problems with that.
  6. First and foremost, Pillar III Note 4.10 shows that BHD 106.5 million is unsecured.  The key point here is that if a loan is unsecured, then by definition it has no collateral.  Consequently, unsecured loans should be excluded from measures of collateral coverage.  That of course would apparently make the ratio higher at 148%.  Note that the data for “Other” and “Unsecured” is switched in both the Arabic and English versions of the 2015 AR as evidenced by data in prior years’ annual reports.   
  7. Second, unless this collateral is pledged to all secured facilities the 148% coverage ratio is meaningless.
  8. As a concrete example, suppose you take a $1mm loan from Bank Arqala (“BA”), assuming you can pass our stringent credit process, and pledge The Trump Tower in NYC.   Our fine bank has collateral worth let’s say a $1 billion, if not multiples more.  But TTTNYC only secures your loan.  If BA makes loans totaling $999 million to other borrowers, the average collateral on the portfolio is not 100%.  One loan for $1 million is over collateralized.  It’s also important to remember that BA like any other bank can only collect what you owe on your loan when it sells (realizes) the collateral you pledged.   If you owe $1.1 million in P&I, BA can take and sell the Trump Tower but only keep $1.1 million not the untold billions TTTNYC is really worth.  The bank must return the rest of the proceeds from the collateral realization to you.  
Collateralization Levels

2015
2014
2013
2012
2011
Unsecured
106.5
105.2
55.9
47.9
40.3
Total Gross Exposure
408.7
361.8
306.0
286.3
217.3
% TGE
26.1%
29.1%
18.3%
16.7%
18.5%






Partially Secured
25.0
17.9
9.9
10.8
12.2
% TGE
6.1%
4.9%
3.2%
3.8%
5.6%






% TGE –All Unsecured
32.2%
34.0%
21.5%
20.5%
24.1%

Comments on collateralization levels: 
  1. As is clear, KHCB has been expanding its unsecured portfolio.  It was at 18.5% in 2011 and 26.1% in 2015.  However, percentages don't convey the full extent of the change.  Unsecured loans have gone from BHD 40.3 million to BHD 106.5 million.  Why is that important?  Because over the period capital has not increased by 2.5x.
  2. In general, but not always, uncollateralized lending is riskier than collateralized.  Tenor (maturity) of the loans would also affect risk.  That info is not available. 
  3. Note the partially secured amounts are net after deducting the partial collateral shown in Pillar III Note 4.10 for “cash” and “other” collateral at face value (no haircut by AA).   
Types of Collateral: 
  1. Pillar III Note 4.10 discloses that 85% of KHCB’s collateral is real estate. 
  2. Three observations.
  3. First, real estate is illiquid and may be hard to sell, unless one is holding a security interest in a piece of land in central Tokyo.  There’s probably not comparable land in Bahrain.  Additionally, it’s well known that with the right nautical partners one can create perfectly good land in Bahrain from the sea as GFH and Arcapita might testify.  Cash or marketable securities would be much more liquid and provide much more protection.  That’s why KHCB advances on average only 60 fils against each dinar of real estate collateral.
  4. Second, real estate is interest sensitive.  As market rates rise so do cap rates (the interest rates used to determine the value of the property). As a consequence, property values decline.  That could be a problem as the US raises interest rates and currency pegged-Bahrain follows along.  The main consequence is likely to be eroding collateral values. Borrowers are shielded from rising rates by the fixed rates on their existing loans--absent covenants in the loan agreements that allow the bank to raise borrower’s interest rates.  Thus, borrowers’ ability to repay should not be affected directly, though general interest rate could indirectly stress their ability to repay by reducing overall economic activity in the Kingdom.   Potential buyers of “seized” collateral are likely to require financing.  If rates on new loans are higher than currently, the sale of real estate will be more difficult, perhaps requiring KHCB to make loan term concessions or reduce the price of seized property being sold.  That leads nicely into the next point. 
  5. Third, when looking at an institution’s exposure to real estate risk, one needs to look at more than the purpose of the loan or the business of the borrower. If a bank makes a loan for a new airline, but secures it with real estate, it has an indirect exposure to real estate along with the direct exposure to the airline.  If the loan is being made because the real estate collateral “compensates” for shortcomings in the airline’s creditworthiness, then the exposure to real estate is more “direct”.
  6. Some 55% of KHCB’s exposure (based on outstanding exposure not collateral values) is secured by real estate, which tempers KHCB’s management’s statement at the bottom of the chart in Pillar III Note 4.3.6 page 88: “The Board approved internal cap for real estate exposure at 40% of total assets. The Bank’s real estate exposure as of 31 December 2015 and 2014 are within the policy limit.”   Not exactly.  Assuming collateral is taken because it is needed to support extension of credit, then KHCB’s real estate exposure is larger than 40%. 
All in all a poor fit for GFH’s announced strategy.

Wall Street Titan "Baffled" by Government Form

Steven T. Mnuchin, Nominee for Secretary US Treasury

When asked to explain how he left $100 million in assets and some "other" information off a required government disclosure form, Steven T. Mnuchin, President Trump's nominee for US Treasury Secretary, stated: “I think as you all can appreciate, filling out these government forms is quite complicated.” 
Once again it's the fault of "Big Government" overreach burdening a self-reliant and hard-working job creator.
Now some or you out there might think that a guy who was a partner at Goldman Sachs and a supremo at several hedge funds would have little problem with the mechanics (note that limitation) of filling out a Federal ethics disclosure form.  But apparently you would be wrong.    
Luckily, the position of Secretary of the Treasury doesn't require a sharp mind or attention to detail --at least in the incoming Administration. 
On a positive note, continued lack of such characteristics could result in a dramatic reduction in the Federal Debt if ten trillion or so was "left out" of the complicated Federal form for reporting national debt.
Danced With the Stars, Still Struggling as an "Apprentice"
The Apprentice appeared to have higher standards, but then all TV is not "reality" TV.

Paging John Kenney Toole.

Russian Showers -- The Grand Cascade

Petergof, Leningrad
An engineering marvel - the fountains operate solely by gravity.  No pumps at all.

Tuesday 17 January 2017

US Presidential Inauguration: Helpful Tips and Tricks


Clearly unsuitable attire for either the formal swearing in or the subsequent celebratory parties.

Friday 6 January 2017

Analyzing KHCB's Financials - Credit Metrics Part 2



Continuing our financial analysis of KHCB's lending portfolio, today we'll take a look at past due loans that have not yet been classified as impaired.

Past Due but Not Impaired Loans (PDNI)
  1. When a bank declares a loan impaired, it takes provisions, ceases accruing income, etc.  It does not do so, with PDNIs.
  2. Often today’s PDNIs turn into tomorrow’s impaired loans.
  3. But note that a PDNI is not of itself conclusive proof that the loan will turn out to be “bad”.  Sometimes a borrower has a legitimate reason for being late and brings the loan current.
  4. What analysts look for is trends in PDNI (increasing, decreasing) and the length a PDNI is past due to form an opinion on future impairments.
Let’s start with a macro view.  What has been the trend at KHCB between 2011 and 2015? 

KHCB Past Due But Not Impaired Loans 2011-2015

2015
2014
2013
2012
2011
BHD Millions
60.8
35.8
39.0
39.5
18.7
% FA & LA
15.7%
10.6%
13.7%
14.8%
8.8%

  1. At first glance, the trend doesn’t look good.  But let’s dig a bit deeper.  
  2. How are the loans distributed by time past due? The theory being the longer a loan has been past due the weaker the probability of full collection.   

KHCB PDNIs Over 90 Days Past Due

2015
2014
2013
2012
2011
90-180 Days
7.93
1.01
1.58
3.12
1.95
% All PDNIs
13.0%
2.8%
4.0%
7.9%
10.4%






180+ Days
17.2
1.0
1.8
4.7
8.3
% All PDNIs
28.3%
2.8%
4.5%
12.0%
44.4%






All Over 90 Days
25.2
2.0
3.3
7.9
10.3
% All PDNIs
41.4%
5.7%
8.6%
19.9%
54.8%

  1. This is an even bleaker picture. 
  2. In 2015 41.4% of all PDNIs are over 90 days past due. 
  3. 28.3% of all PDNIs are over 180 days.  That is PDNIs are skewed to the longest shown past due “maturity bucket”.
  4. The last time the numbers were comparable (2011) there was a massive loss two years later due to provisioning.
  5. AA wonders and maybe you do too how loans can be past due for more than 180 days without being considered impaired.
  6. But can think of two explanations.
  7. The loans are collateralized by highly liquid assets with a robust margin of safety and based on impeccable legally enforceable agreements.  That is, if the borrower doesn't repay, the collateral is sufficient to repay principal and interest in full and the agreement is so watertight that no judicial delay will occur.
  8. Another is something a GCC national banker once told me about “Islamic” banking.   “AA”, he said, “it’s often said that Islam has no miracles.  No wine into water.  No raising of the dead.  No healing of lepers.  And that is true, except with regard to ‘Islamic’ banking.  There halal profit rates miraculously track interest rates even with the latter are at a level so low that the average suk merchant couldn’t make a living.  But that’s not the only miracle. One can make a profit by buying and selling to oneself.” 
  9. AA supposes that if this is true, then loans can be past due 180 days without being impaired.    أعلم الله
So far we haven't found a compelling reason why KHCB would be a good fit with GFHFG's new strategy.

One more post--a look at collateral--and AA's deep dive into KHCB's financials will come to an end.  Perhaps the "gold" is here that bolsters KHCB's credit portfolio.  

Wednesday 4 January 2017

Analyzing KHCB's Financials -- Credit Metrics Part 1

Choose Your P's or C's
Today we'll turn our attention to the heart of KHCB's business--its lending and leasing activity.

Provision Coverage

KHCB Provisions – Amounts in Millions of BHD

3Q16
2015
2014
2013
2012
2011
FA
358.7
318.7
295.8
253.6
245.7
201.6
Provisions
17.3
13.8
13.7
16.3
16.4
18.1
% FA
4.6%
4.1%
4.4%
6.0%
6.3%
8.2%

  1. Provision percentage is calculated on gross portfolio.  Provisions/ (Net FA + Provisions). 
  2. Note:  KHCB does not calculate provisions on its Leasing Assets, but rather measures the equivalent of “credit” risk in terms of impairment to value of the asset.  This is based on the assumption that KHCB can repossess the asset from a defaulting borrower and lease it to another with no loss of principal.  But note that bank management may have more discretion in determining impairment of a LA than loan impairment with a FA.
  3. Over the period KHCB’s FA provisions have been declining as a percentage of outstanding FA.
  4. If the credit quality of the portfolio is high, then a lower provision makes sense. 
  5. If the quality appears to be less than robust (another euphemism), then perhaps the decline in provision coverage is more a matter of earnings management than credit management.  Lower provisions equal higher net income all other things being unchanged. 
Let’s dig a bit deeper to examine the credit quality of KHCB’s portfolio to see if we can form a view on which of these two is more likely the cause. 

Renegotiated Loans
  1. As we discussed in reviewing ADCB, a timely renegotiation of a loan could prevent it from becoming past due and potentially impaired.  Thus, renegotiation can be a tool to manage earnings by lowering provisions as well as to legitimately restructure loans.  
  2. As KHCB states in its 2015 AR, Note 31 Credit Risk page 61:  Exposures classified as neither past due nor impaired financing facilities include facilities renegotiated during the year amounting to BD 49,276 thousand (2014: BD 32,910 thousand) that would otherwise be past due as per their original repayment terms. Italics are AA’s. 
So what has been the trend over the 2011 to 2015 period?

KHCB Renegotiated Loans -  Millions of BHD

2015
2014
2013
2012
2011
Amount
49.3
32.9
17.5
54.4
41.2
%  LA & FA
12.7%
9.7%
6.1%
20.4%
19.3%

  1. There has been a consistent renegotiation of significant percentage of the portfolio each year, except in 2013 when KHCB took significant provisions.   
  2. Side note the portfolio increased from BHD 213 million in 2011 to BHD 360 million in 2016. A key question and one not answered by available information is how many of these renegotiations, if any, are for loans that were in the portfolio in 2011 and how many reflect newer loans?  The answer to which would provide indications on whether KHCB’s underwriting standards were or are lax or there were persistent depressed economic conditions in Bahrain.  Another question would be if the same loans have been renegotiated more than once. 
  3. On that score, if you’re reading along with AA in KHCB’s 2015 AR, a side note.  AA believes the following sentence in the Renegotiated Loans note has been misplaced.  Of the total past due facilities of BD 60,758 thousand (2014: BD 35,785 thousand) only instalments of BD 30,204 thousand (2014: BD 14,108 thousand) are past due as at 31 December 2015”.  That is, it does not refer to renegotiated loans but to all loans and is no doubt an attempt to put a good face on things.  What it does say is that roughly 50% of the payments had been missed on 2015 PDNIs and 40% in 2014.  I don’t know about you but AA isn’t comforted by that information.  That level of past dues raises serious questions why these loans are not impaired.  More on that to follow.  
From the info we do have it looks like KHCB has ongoing weakness in its portfolio.
  
In following posts, we'll look at other portfolio credit metrics to try to determine if there are other signs of weakness.