Friday 10 November 2017

Review of QCB Data on Net Foreign Assets Position of Qatar Banking Sector - Part 2

Hope You're Enjoying the Deep End of the Pool 
Hold Your Breath There's a Lot More to Come
Earlier posts here and here..

FOREIGN ASSETS

Before we dive deeper, a reminder that there are foreign assets that can offset foreign liabilities. 

The question is how much of an offset and when they will be available to offset.


Qatar Commercial Banks
TFA as Percent of TA
EOP
TFA
TA
NFA
Dec-16 273 1,272 21%
May-17 268 1,323 20%
Jun-17 260 1,316 20%
Jul-17 249 1,313 19%
Aug-17 235 1,328 18%
Sep-17 234 1,347 17%


What are the FA composed of?


Qatar Commercial Banks FA
Billions of QAR
Dec-16
Sep-17
Net Chg
USD
Cash
2
4
3
$1
DFB
115
71
-44
-$12
Credit 
95
95
-1
$0
Invest
59
61
3
$1
Other 
4
4
0
$0
Total
274
234
-39
-$11


  1. As expected Qatari banks have drawn down the most liquid portion of their foreign assets, DFB (Due from Banks).  Credit (Loans) and Investments are less easy to liquidate and typically earn higher returns which would make them resources of last choice. 
  2. A key question is how much more room there is to reduce DFB.  We’ll look at that below. 
  3. Note highly liquid FA are some 32% of TFA as of September.  If other FA cannot be liquidated quickly, that would make the adjusted NFA position a deficit of USD 81 billion equivalent.
Who owns these assets?

Looking at Tables 21 and 22 (MMB) or Tables 22 and 23 (QSB), we see that foreign banks operating in Qatar have a negligible share of FA and FL, roughly 1%-2%.  If you’re wondering, the non-Arab foreign banks hold the bulk of FA and FL in this category. That means the burden of NFA falls on Qatari banks.  For ease of computation I have used the TFA and TFL as reported by the QCB to analyze local banks’ NFA, ignoring the de minimis position of foreign banks.


Qatar Banking Sector  - NFA
Billions of QAR
As of 31 December 2016
FA
FL
NFA
Qatar Banks
269
441
(172)
Foreign
5
6
(1)
Total
274
447
(173)
As of 30 September 2017
FA
FL
NFA
Qatar Banks
231 361 (130)
Foreign
3 7 (3)
Total
234 368 (133)

Turning to the domestic banking sector, we see that commercial banks hold the bulk of FA and are responsible for the bulk of FL.  No surprise here, QNB dominates the banking scene, especially in cross-border activity.  Also it appears that the reduction in the overall NFA has been concentrated in the “traditional” banking sector.  Also notice that while the "Islamic" banks have reduced FA and FL, their NFA position has not improved. Of course, QAR 24 billion (USD 6.3 billion) is rather small change indeed for the GOC


Billions of QAR
As of 31 December 2016
FA
FL
NFA
Traditional
228
378
(150)
Islamic
41
64
(23)
QDB
0
0
0
Total
269
441
(172)
As of 30 September 2017
FA
FL
NFA
Traditional
202
308
(106)
Islamic
29
53
(24)
QDB
0
0
0
Total
231
361
(130)


Who owes the Qatari banks? 
Or in other words, where are the obligors?  That’s important for a couple of reasons. First, such information might give an insight into the creditworthiness of these obligors which would in turn help frame estimates of liquidity of FA. If we find that FA are concentrated in one region, then this may have implications for the banks' ongoing business.  If a substantial portion of FA are in other GCC states, then assuming no resolution to the crisis, this business may be ultimately lost to Qatari banks.  As well regional FA concentrations might well make Qatar vulnerable to actions to obstruct the realization of the proceeds of those assets.   Unfortunately, QCB data doesn’t include that information.
However, there is information in the QCB’s Banks Monthly Statements (BMS) reports on deposit transactions between Qatari banks and their overseas offices.  As outlined above, foreign banks operating in Qatar have a minuscule share of foreign assets and liabilities.  So we can ignore their positions and can use the totals here as proxies for the local Qatari banks’ position.  This will be particularly important in looking at foreign liabilities as outlined below.

Qatar Commercial Banks Foreign DFB
Billions of QAR
EOP
DFB
InterBr
% DFB
Dec-16
114
32
28%
May-17
102
34
33%
Jun-17
94
34
36%
Jul-17
84
37
44%
Aug-17
71
36
52%
Sep-17
71
35
50%


  1. FA “Due from Banks” (DFB) held with overseas branches (InterBr) have increased slightly from December in amount and now account for more than 50% of all FA DFB. 
  2. That suggests that there is little room further extensive reductions in foreign asset DFB. The reasons are twofold:  regulatory and business.  Qatari banks will need to maintain foreign asset liquidity to meet prudential ratios.  Withdrawing funds from overseas branches would likely constrain those entities' ability to conduct operations.

FOREIGN LIABILITIES
As noted earlier, the Qatari banks contractually owe the total amount of foreign liabilities not the NFA (net foreign assets) position. 
What are the foreign liabilities?
Qatar Commercial Banks FL
Billions of QAR
Dec-16
Sep-17
Net Chg
USD
NRCD
183
143
-40
-$11
Banks
208
169
-39
-$11
Debt
49
51
2
$1
Other
6
4
-2
-$1
Total
447
368
-79
-$22



  1. Almost 90% of FL is in Non-Resident Customer Deposits (“NRCD”) and Due to Banks (“DTB”). Liabilities one would expect to have relatively short tenors.  This illustrates the vulnerability.  As seen above, some 32% of FA are liquid, while some 90% of FL are liquid.  So there is a fundamental “maturity transformation” mismatch.  Beyond that don’t forget that FL (QAR 368 billion) are some QAR 133 billion greater than FA. 
  2. Note that NRCD have proven to be as volatile as DTB. 
  3. What’s interesting is that NRCD withdrawals been concentrated in private sector deposits.  Foreign non-financial corporate and public sector deposits have been much more stable, though as per QCB’s MBS report these are primarily time deposits.  Thus, it may be that these non-private sector depositors are contractually unable to withdraw funds at present. If this is correct and these depositors want to exit, then withdrawals may accelerate in these categories, particularly if depositors are from the GCC 3+1.
  4. Two data points (August and September) are not necessarily an irreversible trend, but the pace of withdrawals has slowed. Only USD 1.7 billion in NRCD were withdrawn in September, of which USD 1.4 billion were private sector deposits, USD 0.2 billion in non-bank financial institution deposits, and  USD 0.1 billion in public sector deposits. 
  5. DTBs roughly held steady.  New deposits by overseas branches of roughly USD 0.742 billion partially offset withdrawals of USD 0.824 billion by other banks. 
  6. Tracking future developments in these liabilities will give us the best insight into the extent of capital flight in the Qatari banking sector.  QCB BMS reports are available monthly with a lag.
Who are the providers of funding?
As noted above, QCB released data doesn’t provide a geographical allocation of liabilities. 
But there is mirror data on transactions with overseas branches.
Qatar Commercial Banks Foreign DTB
Billions of QAR

DTB
InterBr
% DTB
Dec-16
208
87
42%
May-17
235
122
52%
Jun-17
193
90
47%
Jul-17
180
76
42%
Aug-17
170
62
37%
Sep-17
170
65
38%

  1. Qatar banks have substantial liabilities to their overseas branches.    
  2. Surprisingly, repayment of DTBs has been primarily in intrabank deposits, that is, to overseas branches of Qatari banks. 
  3. From December 2016 to September 2017, Qatari banks have repaid deposits to their overseas branches of some QAR 22 billion (58% of all repayments) and from May to September, QAR 57 billion (88%).   Presumably, this reflects funding pressure on these branches rather than those branches making adverse decisions on the creditworthiness of their parents. 
  4. September marked a change as overseas branches deposited QAR 3 billion (USD 0.742 billion) while other foreign banks withdrew a slightly larger amount equivalent to USD 0.824 billion.
  5. Surprisingly, other DTBs have shown less volatility. This is perhaps because QNB operates major treasury and trading centers outside of Qatar.  And, thus, overseas branch reduction of DTBs is an indirect reflection of creditor flight at those branches. 

At this point, we have taken our “diagnosis” as far as we can with QCB publicly available data.

What further information do we still need? 
  1. A geographical allocation of FA and FL. 
  2. Disaggregation of the aggregate NFA position to identify the positions of individual banks.  Which banks are most at risk?  Which banks have positive NFA and which negative NFA?  This is important because it is unlikely that NFA-positive banks are going to liquidate assets to support NFA-negative banks.      
I'll cover that in a series of posts to follow.  

Review of QCB Data on Net Foreign Assets Position of Qatar Banking Sector - Part 1

Join AA as We Dive Deep into the QCB Data  Be Sure You've Got an Extra Tank

Introduction
Since the GCC 3+1 (Egypt) announced their “blockade” against Qatar, there has been a lot written about the negative economic impact on Qatar, some of it true and some of it seizing on negative news to invent imagined foreign policy successes.  For the latter read Gulf News. 
One point that has been emphasized is that Qatari banks have a negative net foreign asset (NFA) position.  On an aggregate basis Qatar’s banks are borrowing offshore to fund onshore assets.  This is a real vulnerability.
But …
Aggregate statistics can identify symptoms of diseases but further diagnosis is required to determine if the patient actually has a disease and how severe the disease is.  At times aggregate statistics can obscure diseases.  For example, if we assume that Qatar banks had a perfectly balanced position with FA = FL, aggregate statistics would suggest everything is fine.  But if the FA were in shares of GFH and Dana Gas to choose two sterling investments, there would likely be a real problem.  Or one bank had a positive NFA and the other a negative NFA, the potential demand for repayment would be the negative NFA of the second bank, not the net of their two positions as in the QCB NFA data.
What’s important then is taking a detailed look after parking any partisanship in the dispute.
This post will do just that. 
As noted in my introductory post, I’ll be looking at this topic using two main data sources:  QCB statistics and bank consolidated financial reports.  This post will focus on QCB reported NFA. 
QATAR CENTRAL BANK DATA
We’ll begin by looking at data published by the QCB in its “Monthly Monetary Bulletin” (MMB) and ”Quarterly Statistical Bulletin” (QSB).
To set the stage a discussion of aggregate statistics  
Qatar Commercial Banks NFA
Billions of QAR
EOP
TFA
TFL
NFA
Dec-16 273 447 -174
May-17 268 471 -203
Jun-17 260 415 -155
Jul-17 249 388 -140
Aug-17 235 371 -136
Sep-17 234 368 -133
Net Chg -39 -79 40

Qatar Commercial Banks NFA
Billions of USD
EOP
TFA
TFL
NFA
Dec-16 $75.1 $122.7 -$47.7
May-17 $73.6 $129.4 -$55.9
Jun-17 $71.5 $114.1 -$42.6
Jul-17 $68.3 $106.7 -$38.4
Aug-17 $64.5 $102.0 -$37.5
Sep-17 $64.3 $101.0 -$36.6
Net Chg -$10.7 -$21.8 -$11.0

Technical Notes:
  1. QAR amounts rounded to nearest billion.  USD equivalents rounded to nearest hundred million.
  2. Conversion to USD equivalent at QAR 3.64 = USD 1.00.
  3. Note that “Qatar Commercial Banks” includes all banks licensed to operate in the emirate including foreign banks. 
  4. FA = Foreign Assets. FL = Foreign Liabilities. NFA = FA – FL.  T= Total, i.e., TFL = Total Foreign Liabilities.
  5. Sources: Tables 3, 21, and 22 in the Monthly Monetary Bulletin and Tables 4, 22, and 23 in the Quarterly Statistical Bulletin.  The MMB is not published at Quarter end.
Comments:
  1. The aggregate (note that word) NFA assets position of the Qatari banking sector seems rather modest in terms of the resources that the GOQ can bring to bear.
  2. From FYE 2016 through September 2017 Qatar’s banks have reduced their negative NFA position by some USD 11 billion or roughly 22%.  If measured from May 2017, the reduction is about USD 19 billion or 33%. 
  3. During this period, USD 11 billion equivalent in FA have been liquidated to meet repayments of FL and FL have been reduced the equivalent of USD 22 billion.  If measured from May, FA have been reduced some USD 9 billion and FL about USD 28 billion.
At first glance, the banks and GOQ seem to be handling matters well.  The pace of creditor flight in August and September appears to have slowed. But these two data points do not establish an irreversible trend.
Hopefully, there is more than one reader out there (that’s meant in more than one sense) who is thinking and, thus, thinking of saying: “Wait a minute, AA, the Qatar banks don’t owe the Net FA position, they owe the total of foreign liabilities, some USD 101 billion.  That’s a much bigger amount and would strain the GOQ’s resources a lot more.  That’s the measure to use.” 
Keep this straw man’s comment in mind as it will be a recurring theme in what follows.
Let’s turn to that topic and look at Total Foreign Liabilities (TFL) as a percentage of Total Liabilities and Equity (TLE).
Qatar Commercial Banks
TFL as Percent of TLE
EOP
TFL
TL
%
Dec-16 447 1,272 35%
May-17 471 1,323 36%
Jun-17 415 1,316 32%
Jul-17 388 1,313 30%
Aug-17 371 1,328 28%
Sep-17 368 1,347 27%

In light of the negative NFA position, these ratios are concerning.  More importantly they understate the relative importance of TFL as a percent of market funds.   To correct this let’s remove equity from the denominator.
Qatar Commercial Banks
TFL as Percent of TL  (No Equity)
EOP
TFL
TL
%
Dec-16 447 1,093 41%
May-17 471 1,131 42%
Jun-17 415 1,124 37%
Jul-17 388 1,119 35%
Aug-17 371 1,134 33%
Sep-17 368 1,153 32%

  1. Note:  In adjusting TLE for equity I deducted QAR 25 billion in perpetual subordinated debt which is Tier 1 Capital, but is not accounted as capital in the QCB statistics.
  2. TL also includes “Other Liabilities” which average about QAR 4 billion.  However excluding these does not materially change the percentages. So I have not.
  3. The ratios are higher.  So even more cause for concern. NPA fairly dramatic picture of dependency on foreign funding. 
While this sharpens the analysis, it doesn’t provide a complete answer.
Diagnostic Questions
To get that answer ideally we need to answer at least the following questions.
Foreign Assets 
  1. What are the foreign assets composed of? For example, if the banks foreign assets are substantially in liquid high credit quality instruments, then the absolute net additional funds they require to settle all foreign liabilities are much smaller, approaching the NFA position. 
  2. Which banks in Qatar own the assets?  Because the QCB statistics are for the entire Qatari banking sector, they include foreign banks operating in Qatar.  How big is their share of the FA?  List of licensed banks in Qatar.
Foreign Liabilities
  1. What are the foreign liabilities composed of?  How much is bank-provided funding (typically expected to be more volatile), customer deposits, and debt instruments?
  2. Which banks, local or foreign, are responsible for paying these liabilities?  How does this compare to their shares of FA?  For example, if foreign banks have borrowed foreign currency to fund their QAR denominated local operations and this is a significant amount of total FL, then the problem for Qatar is less because these banks are unlikely to be affected by creditor flight.  And if it does occur, then their parents are likely to step up.  But, if this is the case, then which foreign banks hold the large FL and do they have access to funding?   There’s a difference between HSBC, Stan Chart, and BNP on the one hand and Bank Saderat on the other, though we might reasonably expect Bank Saderat to be a “samak saghir” in Qatar.
  3. To whom is the money owed?   For example, if much of the TFL funding is intercompany, it is likely to be more “sticky”, unless the overseas units are under pressure themselves  On the other hand, if a significant portion of the TL are owed to GCC3+1 entities, then the problem is greater.
Additional Foreign Assets and Liabilities
  1. Are there foreign assets and liabilities of the Qatari banks that are not recorded in the QCB data? These are most likely those of subsidiaries.  If contagion spreads to foreign subsidiaries, then the Qatari parents may have additional foreign currency burdens, unless they “walk away” from these subsidiaries.  On the other hand, a foreign subsidiary may be induced to make a timely deposit to help out a parent.  
Individual Bank Positions
  1. As noted earlier, what are the individual positions of local Qatari banks?  Qatar Bank A which has a positive NFA position is highly unlikely going to bail out Qatar Bank B which has a negative one.  Therefore, we need to disaggregate the local banks’ TFA and TFL to individual banks.
We can answer some but not all of these questions from the QCB data. And will start doing that in the next post.

Thursday 9 November 2017

Qatar Banking Sector - How to Measure the Net Foreign Assets Position

AA Absolutely Loves Quotes Like This Though It's Hard to Match "Live Your Life" 

This is the first of four planned posts on Qatari banks’ vulnerability to cross-border foreign currency exposure in light of the measures taken by Saudi Arabia, the UAE, Bahrain and Egypt (the GCC 3 +1). 
Before getting into a detailed discussion of the numbers, it’s important to ensure that we’re measuring cross-border foreign currency exposure vulnerability properly. 
Currently analysts and the press are focused on Qatari banks’ Net Foreign Assets (NFA) position as reported by the Qatar Central Bank (QCB).  It’s a “convenient” measure:  a single number available monthly.   But sadly reality is rarely, if ever, that neat. 
To be very clear, I don’t think there is a fault in the QCB’s NFA statistics, but rather that they give only a partial picture of Qatari banks’ positions and vulnerability.
Why?
The statistics are an aggregate of the separate foreign currency positions of all the banks operating in Qatar with non-resident counterparties.
What does that mean? 
  1. Net Positions Obscure Gross Positions: Assume a country whose banking sector is composed of only two banks, Bank A and Bank B.  Assume Bank A has borrowed local currency to fund foreign assets of some USD 55 billion and Bank B has borrowed foreign currency to fund local currency assets of USD 55 billion.  According to QCB’s valid methodology, NFA = 0, an apparently “happy” cross-border position.  But if foreign creditors cease lending to the country, Bank B needs to pay USD 55 billion to them. If it can’t, the central bank needs to step up unless one assumes the highly unlikely event that Bank A voluntarily liquidates its foreign assets and makes the resulting foreign currency available to Bank B.   The situation is even more complicated when as usual there are multiple banks in a country.  The net of all their individual positions obscures the position of each bank.  And it is the position of each bank that is critical to the ultimate demand for foreign currency on a country.
  2. Net Positions Obscure Exposure to Specific Creditors:  When assessing creditor flight, it’s important to know where the creditors are.  In the case of Qatar, creditors in the GCC 3 +1 are most likely to take “flight”.  Other creditors less so because the crisis is at present driven by political not economic/credit issues. If exposure to the GCC 3 +1 countries is large, then the potential funding gap from their withdrawal is large.  If not, the problem may be minor. 
  3. The standard NFA statistics implicitly assume that assets and liabilities are equally liquid. They “ignore” the fact that a bank’s basic business is maturity transformation – typically borrowing short and lending/investing long. On a contractual basis a good portion of assets (likely a majority) will almost certainly have maturities far in excess of liabilities.  If assets cannot be liquidated quickly because of contractual maturities, the central bank (here QCB) may have to provide additional foreign currency funding to “bridge” the maturity gap. That will be an immediate call upon the central bank’s foreign assets.  If the banks’ assets cannot be realized at par, the parent or central bank may have to provide additional funds to make up the shortfall from asset realization.  If, for example, Qatari banks hold their foreign assets in US Treasury Bills, realization at par is almost certain.  If they hold Dana Gas or GFH shares, less so, much less so.
What’s needed then is additional information on Qatari banks’ foreign currency exposure that supplements the QCB reported NFA position by factoring in the cross border foreign assets and liabilities of local Qatar banks that are not included in the NFA position, subtracting the position of overseas banks operating in Qatar (if their position is meaningful), disaggregating “net” positions into gross Foreign Assets (FA) and Foreign Liabilities (FL) by region (ideally country), and by individual Qatari bank, as well as understanding the liquidity of gross FA of individual banks.  That requires looking at the foreign asset classes the banks hold. 
We can derive some of this information from other data published by the QCB.
But we are still left with unanswered questions.
To answer those questions, we can turn to the fiscal year end consolidated financial reports of Qatari banks, though this method is admittedly imperfect. 
In the third post I’ll discuss the theoretical limitations as well as  practical constraints of using this method.  And why I chose this method as an analytical tool despite those drawbacks.  For example, not all Qatari banks provide a geographical allocation of assets and liabilities. Accounting standards do not require the provision of this data for “parent only” statements that would eliminate a key issue with using consolidated statements. That is Qatari parents have less direct access to the foreign assets of their overseas subsidiaries than to those of their foreign branches.  Also the parents are not legally liable for their subsidiaries' foreign liabilities absent having provided a guarantee. 
One final point. 
This exercise will result in estimates of Qatari banks’ vulnerability—hopefully more comprehensive than the NFA reported by QCB.  But note that both are estimates not “hard” numbers. Vulnerability is composed of two elements: structural position and market sentiment.  We can get a reasonable “fix” on the aggregate structural position, but not an exact number due to the absence of sufficiently detailed information. Pricing and timing of realization of assets remain uncertain and dependent on many factors.  Creditor sentiment—responsible for the “flight” in capital flight--is harder to specify than the structural position.  We  see creditor or market sentiment most clearly post-facto.    All these factors are uncertain and inter-react making this a complex system.  As such, it’s not capable of being modelled accurately.  Therefore, the final post will look at some scenarios to set a range of possibilities and hopefully spark some thinking by others as to how to refine the analysis.
The following three posts will detail the results of an analysis of:
  1. QCB data
  2. Data from individual Qatari banks’  audited annual consolidated financial statements
  3. Various FX exposure scenarios 
These posts will be typical-AA excessively-detailed journeys through the data.  To whet your appetite for the long journey that follows, three teasers.  Note that the consolidated financial statement data cited below is as of 31 December, the only time that banks are required to publish detailed risk management notes.
The Qatar banking sector’s estimated foreign asset and liability exposure to other GCC states (OGCC) is minor in the context of its total foreign asset (TFA) and total foreign liability (TFL) position as of 31 December 2016.
  1. Based on a sample of  7 Qatari banks consolidated financial statements, the base worst case exposure to other GCC states (OGCC)—assuming no realization of any FA held in OGCC—is estimated at some QAR 89 billion (the aggregate of all banks TFL  to OGCC) or some USD 24.5 billion. 
  2. If we assume that Qatari banks will be able to realize their FA at par value, and look only at those banks with a negative NFA, the exposure is QAR 31 billion equivalent to USD 8.5 billion. 
  3. If we take the aggregate position of all Qatari banks, the aggregate NFA is some QAR 18 billion or USD 4.9 billion.  For the reasons outlined above AA thinks this scenario is unlikely.
  4. While we are missing detailed data for four banks, it’s unlikely that they will dramatically change these amounts.  AlKhaliji Qatar has disclosed overseas subsidiaries with gross assets of QAR 8.9 billion (USD 2.4 billion), primarily in the UAE.  QIIB and IBQ do not appear to have foreign branches or operating subsidiaries overseas.  Qatar Development Bank, the missing fourth bank, doesn’t release financials but from the QCB data we can see that it is funded primarily with equity and has a positive NFA of some USD 84 million equivalent. 
  5. If we adjust the base worst case scenario for AlKhaliji's overseas subsidiaries' total assets of QAR 8 billion, the result is QAR 97 billion or USD 26.6 billion.   
An estimate of the Qatari banking sector’s aggregate NFA using individual bank consolidated financial reports results in a much lower aggregate NFA than QCB figures as of 31 December 2016. 
  1. As per consolidated financials, the aggregate NFA is some USD 25 billion equivalent versus USD 48 billion equivalent from QCB data as of 31 December 2016. 
  2. While the consolidated financial “result” is based on a sample and thus does not include all Qatari banks, I think it’s unlikely that the remaining banks could generate an additional negative NFA equaling USD 24 billion equivalent.  Why?  Because the total liabilities (foreign and domestic) of these banks ex-QDB equal QAR 112 billion (some USD 31 billion). It’s highly unlikely that they have foreign liabilities equal to 74% of total liabilities. As well there are likely to have some FA to offset their FL.  
Qatari banks with the aid of the Government of Qatar (GOQ) should be easily able to meet the challenge posed by the GCC 3 +1’s actions. 
  1. However, if the GCC 3 +1 can create a situation in which other foreign creditors withdraw funding support, the burden on the banks and the GOQ would be substantially higher.
  2. Using QCB data the worst case would be aggregate TFL of some USD 101 billion equivalent as of September 2017.  Consolidated financial report data after 31 December 2006 isn’t available for comparison.  But if we use the “spread” at 31 December 2016 between QCB’s computation of NFA and that derived from bank financial reports, the estimated consolidated finance report-derived NFA would be some USD 9 billion larger, i.e. USD 110 billion.