FINA6282 Economics for Finance Practice

Blackstone and the Sale of Citigroup’s Loan Portfolio

Group 5, Class A
LI Hao 1155118617

LIU Qinyu 1155119131

LIU Can 1155096444

WANG Yuling 1155120493

1. Does this transaction make sense for Blackstone? Why is Citigroup seeking to sell the portfolio of
leveraged loans?
l For Blackstone:
Overall, we believe the transaction makes sense to Blackstone.
Below is our risk-return analysis:
1) Risks:
i. Blackstone needed to bear market risks since the debt provided by Citigroup was required to be
marked-to-market. Recently the financial market had been highly volatile, and it seemed that the
economy was moving into a deep recession. It was hard to predict the future trend of the market. If
the market was still depressed in the following years, and the total value of the loan portfolio fell
below 66.4% of its face value, Citigroup had the right to ask for additional collaterals.
ii. Blackstone also faced operational risks. Default for any individual credit would adversely impacted
Blackstone’s overall return on the portfolio. However, as a debtholder, Blackstone did not have the
right to get involved in managerial and operational decisions.
2) Return:
i. Blackstone had performed certain research and due diligence, which ensures that they understood
the risks in the loan portfolio.
ii. Besides, the partner of Blackstone in purchasing the loan pool, TPG, had also researched and even
participated in many of the LBOs.
The above two points reduce the asymmetric information between the buyer and the seller.
iii. Under current situation, Citigroup had already lost several billion on the leveraged loans they held.
So, Citigroup might be willing to sell the loans far below their fundamental value, which seemed to
be a good deal to Blackstone.
Therefore, Blackstone and TPG definitely have a deeper insight into the underlying LBOs than other investors
and even Citigroup. It was very likely that the loans were undervalued, and Blackstone could make a profit on
the deal.

l For Citigroup:
Its large loan volume was bringing great challenge to Citigroup.
1) Leveraged loans were costly from a regulatory capital perspective since they received a 100% risk weight.
Compared to real estate loans and highly rated securities, leverage loans should provide much higher risk
compensation.
2) Citigroup should meet a capital requirement of 8%, which means, for every billion dollars of leveraged
loans, Citigroup would need $80 million of capital. Citigroup would like to sell the leveraged loans
because as its leveraged loan portfolio reducing, its capital requirement would be easier to fulfill.
3) A large part of these loans was intended to be sold, and their value was marked-to-market under
accounting requirements. Therefore, as leveraged-loan index kept dropping in the last two years,
Citigroup had lost around $3 billion on its leveraged loans from the beginning of 2017 to March 2018.
And Citigroup’s earnings and its stock price were under great uncertainty as the pricing of leveraged
loans still fluctuated in a wide range.
However, in financial world, the question of whether a transaction is truly meaningful for the financial
institutions should not only be qualified but also be quantified. The following part will illustrate the
quantitative methods about how to price the loan, thus to make the judgement.
2. Using the data in the case on default and recovery rates, what is the value of the deal for private equity
firms on a discounted cash flow basis? What is the implied IRR?

A. Discounted Cash Flow

Formula to calculate the discounted cash flow of this deal is shown below

Expected Discounted Cash Flow

Q
789:;:<9 =>?@:89A ∗ =;CDA (FG;HIH>J) + M:NCH:;? M>9: ∗ =;CDA (O:P>GJ9)
= 6
OI<NCG89 M>9:A
RST

, which is calculated based on several adjustments and assumptions given by the information in the case.

1) Marginal Default Probability

W=AVT − W=A
U=AVT =
1 − W=A
, which is shown below

, while Probability of Survival = 1- Marginal Default Probability

2) Interest Payment for Year i

As given in the case, the interest rate is equal to LIBOR+309bps for total 6.11 billion leverage loan.
Thus, yearly interest would be 6110*(LIBOR of that year + 3.09%)

3) Recovery Rate
From three sets of data given in the case, we choose the recovery rate from 1982 to 2004. Reasons are as
followed
i) Economic cycle
U.S.A witnessed two business cycles from 1982 to 2004, recovering from the bottom to the peak
twice. Thus, data from this
period is quite
representative, as we
could view the leverage
loan market in a big
picture, pricing the loan
in a fair way.
 ii) High Tech Bubble
In the early 2000s, the US market also experienced the so-called high-tech bubble, where all the
glories of the tech companies collapsed. Our group
searched the information on the target companies whose
loans are considered by Blackstone, and surprisingly found
that most of the companies were engaged in the internet
related business. Therefore, recovery rate of 1982-2004, a
period that covered the heat, collapse, and recovery of the high-tech bubble, seems to be more
persuasive.

4) Discount Rate
Our group uses CAPM to calculate the discount rate
Discount Rate = M[\[]^_`^a + b ∗ (Mc − M[\dAef]^Ag`h )
Our group uses Forward LIBOR Rate as M[\[]^_`^a , and is given by the case as 0.2.
The market return of the leverage loan is calculated by using the leveraged-loan index (LCDX). As the
index dropped from par level in summer of 2007 to 92 cents on the dollar, the annualized yield of this
o
Tkk p
period is calculated as 4 ∗ j n = 11.27%, being perceived as market return.
lm

However, the risk-free rate included in risk-premium should be calculated by using the historical data.
Since this part is not given in the case, in order to match the investment horizon, our group annualized
the forward LIBOR rate and perceived it as historical risk-free rate. Reasons are as followed: The
purpose of Blackstone to launch this transaction is to buy the dip. Therefore, the risk premium should
include their expectation that the leverage loan market is going to recover instead of keep slumping to
the bottom. Thus, annualized LIBOR could somehow reflect their perception of future expectation.

In this case, the market premium is set at 6.58%, and the annual discount rate = LIBOR + 0.2*6.58%

After all the establishment of assumptions and adjustments, our group calculated the expected
discounted cash flow by using binomial tree (real option). The calculation process is illustrated in the
following graph.
Highlight Point: 1) Nominal amount of $6110 MM would be paid back in the end of Year 5 if no default.
2) Cash Flow of each period should be added back to the last year’s expected cash flow,
and then discounted with the probability of no default of that year.
 Exhibit 1 Expected Cash Flow
Year 1 Loan Unsecured 2 Loan Unsecured 3 Loan Unsecured 4 Loan Unsecured 5 Loan Unsecured
Discount Rate 5.1 9% 5.93% 6.1 9% 6.31 % 6.45%

Default 5.91 % 3551 .44 685.85

Nominal 250.42

No Default 94.09% 425.26 Default 8.1 7% 3551 .44 685.85

Nominal 400.1 2 Nominal 346.31

No Default 91 .83% 470.47 Default 8.04% 3551 .44 685.85

Nominal 432.02 Nominal 340.85

No Default 91 .96% 486.36 Default 7.1 5% 3551 .44 685.85

Nominal 447.23 Nominal 302.93

No Default 92.85% 493.69 Default 5.76% 3551 .44 685.85

Nominal 458.39 Nominal 244.1 2

No Default 94.24% 661 2.24

Nominal 6231 .30
Discounted 5900.1 0 Discounted 5904.53 Discounted 5963.54 Discounted 6029.41 Discounted 6083.25

Expected Cash Inflow 650.55 Expected Cash Inflow 732.33 Expected Cash Inflow 680.90 Expected Cash Inflow 604.87 Expected Cash Inflow 4776.92
Expected Cash Outflow 1 85.55 Expected Cash Outflow 21 3.74 Expected Cash Outflow 223.65 Expected Cash Outflow 228.22 Expected Cash Outflow 4043.55
Total Cash Flow 465.00 Total Cash Flow 51 8.59 Total Cash Flow 457.25 Total Cash Flow 376.65 Total Cash Flow 733.36

Based on the binomial tree above, the discounted cash flow for the $6110 MM is $5900.10 MM, referred as $96.57 per $100.
B. Implied IRR

The calculation of implied IRR is based on the principle of NPV = 0, which is equivalent to PV(Cash
Inflow)=PV(Cash Outflow)
Therefore, we need to calculate the expected cash inflow and outflow of each period.

Cash Inflow of Year i

= (Prob. (Default) ∗ 6110 MM + Prob. (Survival) ∗ Recovery Rate) ∗ DEFG. (HIEJKJLM)NOP
Cash Outflow of Year i = 3810 MM ∗ (LIBOR + 100bps) ($3810 MM is the nominal amount of recycling
provision borrowed from Blackstone)
Expected Cash Flow = Cash Inflow of Year I – Cash Outflow of Year I

Meanwhile in Year 0, Blackstone paid 1260 MM as capital expenditure, which should be also considered as
cash outflow in Year 0.

As illustrated above, the implied IRR is 27.67%

3. Assess the purchase price using information on CDS spreads.

a) Using historical recovery rates, what is the implied probability of default? What is the implied IRR?

A. Implied probability of default

Our group computes implied default rate by using:
Probability of default in a given year=CDS spread/(1-recovery rate)
(1) CDS spread
Our group selects the most updating data ranging from 3/1/08 to 3/14/08. According to Exhibit 3 and 7,
Alltel, First Data, Flextronics and Harrah’s debts are senior unsecured note, Tribune and TXU’s debts are
term loan.

The CDS spread for those unsecured notes are calculated as average weighted by debt face amounts to
represent the unsecured component in the portfolio. For the secured part, Tribune presents extremely
high CDS spread (about 3 times as high as unsecured notes’) for some reason, thus our group decides to
use TXU’s information alone. After deriving CDS spread for the two components respectively, our group
computes the weighted CDS spread for the whole portfolio: 75%*secured debt CDS spread + 25%*
unsecured debt CDS spread. The result is shown in the following table:
 Table 1: Credit spread calculation

The potential flaw here is that the CDS spread reflects the current market sentiment towards the
upcoming crisis and doesn’t correspond to the recovery process in the future 5 years. The implied default
rate would be upward biased.

(2) Recovery rate

Our group calculates again the average recovery rate weighted by proportion of secured and unsecured
debts, i.e. 75% and 25%. The result is 69.35%.
The implied default rate computed with these above two inputs is 23.57% for each year.

B. Implied IRR
Our group uses the same IRR calculating methods as in Q2 and only changes the default rate to 23.57%. As
observed, the default rate has increased due to higher credit spread and relatively higher recovery rate
which decreases the denominator at the same time. As a result, the net cash flow in Year 5 is negative for 2
reasons: First, as time goes by, the chance to experience no default and get the $6110mn principal back in
Year 5 becomes smaller and smaller, thus the expected cash inflow in Year 5 is lower; Second, Blackstone
is still obligated to pay the final year coupon and $3810mn principal back for Citigroup debt.
Among the cash flow series over 5 years, the initial equity investment outflow and the negative cash flow
in Year 5 outputs a total undiscounted cash flow of $-312mn, therefore it is impossible to find IRR based on
our assumptions and calculations. The calculation process is presented below:

Table 2: Implied IRR calculation

Scenario analysis under good market sentiment

Our group uses average 5-yr CDS spread on Aug-07 to repeat the above calculations. Since the market
hasn’t responded intensively to the upcoming crisis, the implied default rate is 15% under lower credit
spreads and the implied IRR becomes 25.46%.

Table 3: Implied IRR calculation under good market sentiment

b) It is useful to note that buying the loan in combination with a CDS on the loan makes an investor’s
payoffs essentially riskless. What are the annual cash flows from such an investment? How would you
discount those cash flows? What is the value of such an investment?
A. Annual cash flows
Our group make 2 adjustments in the annual cash flows:

First, when the loan portfolio defaults, the CDS seller would guarantee that Blackstone gets its full
notional amount, therefore the recovery rate is 100%; Second, when the loan portfolio doesn’t default in a
given year, Blackstone would pay CDS premium to CDS seller, thus the cash flow from coupon or
principal is deducted against the premium outflow

Discount rate
Since the loan portfolio is protected by CDS, the loan portfolio would be discounted by risk-free rate. Our
group comes up with 2 choices, forward LIBOR and swap curve.

(1) Swap curve is better than LIBOR, because LIBOR only contains interest rate risk, therefore only the
credit risk of the underlying portfolio should be excluded through CDS. LIBOR doesn’t reflect all other
risks compared to asset swap;
(2) Our group tries deriving swap spread by deducting CDS spread from portfolio return computed in Q2.
The problem here is that the portfolio return, which is lower compared to current return in the crisis
period, reflects the future recovery process of the market, while the CDS spread we used reflects the crisis
sentiment. Therefore, the method to use swap curve as benchmark is given up.

Based on the same method as in Q2, the value of this investment is $5381.81mn, equivalent to $88.08 per
$100 debt face value.

4. In which valuation approach do you have more confidence? Why?

The key difference between the first approach in Q2 and the second approach in the first question of Q3
lies in the choice of default probability.
In Q2, our team turns to a historical approach where default probability is based on historical data
provided in Exhibit 6 given in the case. Our team uses cumulative default rate to solve for marginal
default probability, with time span starting from 1981 to 2005. This historical data makes sense. Because in
the early 1980s, there was a major worldwide economic crisis. Specifically, in US, the Fed raised Federal
Funds Rate to double digits to rein in inflation. During that period, the fixed income market was also in
turmoil. Starting from 1982, the U.S.A began to step out of the mud with an extended recovery, jumping
from bottom to the peak. The following decades witnessed the US loan crisis, Asian financial crisis, and
the internet bubble. Since we always have the boom and the slump in economy, the historical data could
somehow demonstrate the fair value of the deal, without considering market sentiment. The calculation
of Q2 is based on the assumption that the market would always recover from recession.

However, from 2005, the economy in U.S was in full expansion. Right after Lehman Brothers fell, the deal
between Citigroup and Blackstone truly happened. The calculation of Q3 is largely a market-based
approach, reflecting market participants’ prospective on the future, as the CDS spread truly reflect the
results of current market transactions. As we could see from the data that the CDS spread has been
widening from early 2007, showing that those large financial institutions sensed the potential risk. Under
such circumstance, the method used in Q3 is more appropriate to price the deal.

5. Based on your calculations, how attractive is this deal to Blackstone?

Mr. Goodman at Blackstone wants to buy the dip since the market participants are pessimistic. But he
thinks the selloff is overdone and market will eventually rebound. So, to him, the offered price 83 cents on
a dollar is a screaming buy. However, the price in Q3 approach is closer to Blackstone’s final offer price,
reflecting the market fear. Since using historical data in Q2, the fair value of the loan is $96.57, if we
believe the portfolio will recover almost to its face value at the end of the 5-year investment horizon. Thus,
our group has more confidence in using the approach in Q2.
Appendix

Exhibit: Value of portfolio with CDS (spread data on 3/1/08-3/14/08)

Year 1 Year 2 Year 3 Year 4 Year 5
3.87% 4.61 % 4.87% 4.99% 5.1 3%

Default Default Default Default Default

prob 23.57% prob 23.57% prob 23.57% prob 23.57% prob 23.57%
CF 61 1 0 CF 61 1 0 CF 61 1 0 CF 61 1 0 CF 61 1 0

No default No default No default No default No default

prob 76.43% prob 76.43% prob 76.43% prob 76.43% prob 76.43%
CF -1 6.1 1 CF 29.1 1 CF 44.99 CF 52.32 CF 61 70.88
PV Yr 2,3,4,5 5445.88 PV Yr 2,3,4,5 5540.46 PV Yr 4,5 5672.86 PV Yr 5 5856.1 1

Expected 5590.09 Expected 5696.93 Expected 581 0.28 Expected 5955.94 Expected 61 56.53

PV in year 1 5381 .81

88.08202

Exhibit: Value of portfolio with CDS (spread data on Aug-07 under better market sentiment)
Year 1 Year 2 Year 3 Year 4 Year 5
3.87% 4.61 % 4.87% 4.99% 5.1 3%

Default Default Default Default Default

prob 1 5.00% prob 1 5.00% prob 1 5.00% prob 1 5.00% prob 1 5.00%
CF 61 1 0 CF 61 1 0 CF 61 1 0 CF 61 1 0 CF 61 1 0

No default No default No default No default No default

prob 85.00% prob 85.00% prob 85.00% prob 85.00% prob 85.00%
CF 1 44.28 CF 1 89.49 CF 205.38 CF 21 2.71 CF 6331 .26
PV Yr 2,3,4,5 5758.00 PV Yr 2,3,4,5 581 8.67 PV Yr 4,5 5895.25 PV Yr 5 5990.74

Expected 5933.44 Expected 6023.44 Expected 61 02.04 Expected 61 89.43 Expected 6298.06

PV in year 1 571 2.37

93.4921 5