Today in LTEV Mysteries

Ok folks, let’s have a competition. According to page 3 of this document, we’re paying €8.5 billion for the first tranche of loans, which are backed by property with a calculated long-term economic value of €10.5 billion. First person to provide full details of how exactly this works gets a copy of the old NAMA protest article signed by all 46 guys. Should be worth a fortune in years to come. Zhou is doing trojan work on this right now and has been installed as odds on favourite by Paddy Power.

38 replies on “Today in LTEV Mysteries”

Surely my highly plausible “why wouldn’t you discount long-term economic value by the discount factor for the equivalent of 3.7 years?” (http://www.ronanlyons.com/2010/03/31/eu-scuppers-long-term-economic-value-as-namas-first-tranche-goes-through/) is in with a shout!!

Big thing to watch is the discount on the remaining tranches. We’ve transferred 50% of the random “associated loans” with this tranche, but only 10% of the land and 10% of the actual property (remember what this was all about?!). The nature of future tranches will be very different, and therefore so will/should the discount.

Not sure of the question – there are 16bn of loans, which NAMA is paying 8.5bn for. These are secured by property with a current market value of 9.44bn, which is adjusted to 10.5bn to account for LTEV. Or am I missing something?

Ok Ronan, sounds like you’re perhaps circling in on the right answer. But full details required to get the grand prize. I left my homework at home doesn’t work as an excuse.

John: The question is what’s the science behind the 2 billion gap between LTEV of properties and the “total consideration” paid.

Need to be careful here as the guys at NASA have identified a new super black hole and in deference to the Irish man made one, have named their’s Nama2.
As far as NASA is concerned, the black hole, indeed any balck hole, sucks in anything within reach and deosn’t care a whit for value, discounted or other wise. Net result is empty political rhetoric. Another vacuous black hole.

How about page 2 of the document which says that we can recover 2-3bn of costs from the banks for chasing debts and due diligence. Maybe the 2bn difference between LEV and consideration is an offset of these costs?

I note that the biggest difference is with Anglo (1.55bn and 25% of LEV)?

Maybe this was dealt with at the NAMA news conference yesterday – I know there was one because the IT carry a picture http://www.irishtimes.com/newspaper/ireland/2010/0331/1224267400475.html
and Brian Lenihan referred to it. Maybe some bright spark asked the question then?

Also under the traditional defintion of haircut (1-LEV/Loan Value) isn’t the haircut 34% not 47%?

@DE
“Would it be that hard for the government to release an explanation of where these numbers come from”
Well, that assumes that there is somewhere a rational explanation. We made them up in a hurry is a non-rational explanation.
I did like today Alan Dukes constant cries on Newstalk that “its a moving target”…this was just before he said that we would be back to the well again, even 22.3b is not enough.

We’ve 3 figures to consider:

€10.5 bn long-term economic value
€ 9.4 bn current market value
€ 8.5 bn transfer value.

Long-term economic value is based on the statutory calculation of the value of the loans.

Current market value is the estimate, validated by third-party advisors.

Transfer value includes associated derivatives and swap agreements. Remember there are €14bn of derivatives sitting beside the €81bn of loans. With short-term interest rates down significantly, there may be considerable losses on these derivative contracts.

NOTE: the third-party valuers were rather aggressive on valuations during the boom-times. Now, they aren’t allowed use professional-liability-insurance and shall be liable for errors in their estimates provided to NAMA. Thus there is a considerable gap between the “blue-sky” long-term economic value and the actual transfer value.

I want this prize!

CMV = Current Market Value of Underlying property as at Nov 2009.
LTEV = Long-term economic value.
T1 = First Tranche.
T1_Portf = Portion of a particular bank’s portfolio comprised in the first tranche.
Consid = Consideration paid.
Total_Enf_DD_Costs = Enforcement costs for entire portfolio which has been front loaded.
Loans = Loans which constitute Bank Assets and are being valued.
Prop = Underlying property on which loans are secured.
Indiv_Prop = Individual Property.
Portf_Prop = Aggregate Property for a Portfolio.
SDR = Standard Discount Rate of 5.25% (5% Enforcement; 0.25% Due Diligence) as specified in paragraph 8 of the Regs.
Enf_DD_Costs = Enforcement and Due Diligence costs calculated by applying the SDR.
CFDR = Cost of Funds Discount Rate as set out in paragraph 2(2) of the Regs.
Portf_Prop_Cap = 20% cap on uplift on aggregate property LTEV in a portfolio (para 9(2) of the Regs).
Indiv_Prop_Cap = 25% cap on uplift on LTEV of individual portfolio (para 9(1) of the Regs).
Regs = National Asset Management Agency (Determination of Long-Term Economic Value of Property and Bank Assets) Regulations 2010.
the Act = NAMA Act 2009

PROCESS

1. CMV of each Indiv_Prop is assessed by valuers.

2. LTEV of each Indiv_Prop is calculated in accordance with the mysterious and secret LTEV_Indiv_Prop model.

3. Indiv_Prop_Cap is applied to each LTEV_Indiv_Prop to cap it at 25% uplift.

4. LTEV_T1_Portf_Prop is calculated by aggregating the LTEV_Indiv_Props for that bank in T1.

5. Portf_Prop_Cap is applied to LTEV_T1_Portf_Prop for each bank to cap it at 20% uplift. This assumes that T1_Portf constitutes an “acquired protfolio” for the purposes of the Act.

6. If LTEV_T1_Portf_Prop uplift is greater than 20% then each LTEV_Indiv_Prop is reduced pro-rata to ensure that LTEV_T1_Portf_Prop uplift is capped at 20%. (I assume this step is taken because it is the only way I can see the LTEV_T1_Portf_Prop being applied as each Indiv_Loan is valued individually with reference to the underlying LTEV_Indiv_Prop.) It is apparent this did not happen for T1 because LTEV_T1_Portf_Prop uplift is less than 20% for each bank.

7. Each LTEV_Indiv_Prop is then fed into the mysterious and secret LTEV_Loan model.

7.1. The appropriate CFDR [Para 7 of the Regs; Sections 79(2)(b)(v), 79(2)(c)(v), and 79(2)(c)(v) of the Act] is also fed in to the model.

7.2 The SDR is also fed in [79(2)(c)(vi) of the Act].

8. The LTEV_loans are then aggregated to give LTEV_T1_Portf_Loans for each bank.

9. SDR is applied to the estimated total NAMA figures for each bank in accordance with the NAMA model to give Total_Enf_DD_Costs for each bank. [I will return to this below].

10. The difference between estimated Total_Enf_DD_Costs for each bank and the T1_Portf_Enf_DD_Costs for each bank is further deducted from the LTEV_T1_Portf_Loans for each bank to give Consid_T1_Portf for each bank.

THE APPLICATION OF THE STANDARD DISCOUNT RATE:

IF we assume that
(T1_Consideration + Total_Enf_DD_Costs) = LTEV_T1_Loans
THEN
LTEV_T1_Loans = 119.7% T1_Prop_CMV
LTEV_T1_Loans = 107.5% LTEV_T1_Prop
Haircut from BookVal_T1_Loans to LTEV_T1_Loan is 29.5%

IF we apply 29.5% Haircut to the entire NAMA loans (€81bn)
AND WE ASSUME that €2,788bn = Total_Enf_DD_Costs for all NAMA loans
THEN
All_Enf_Costs = 4.88% LTEV_All_NAMA_Loans
All_Enf_Costs = 4.65% of (LTEV_All_NAMA_Loans + All_Enf_DD_Costs)

This is 0.85% outside the 5.25% expected.
There are 2 possible explanations:
(a) The frontloaded €2.788bn does not include the Enf_DD_Costs for T1 which have laready been taken account of in the T1 valuations are so are not front loaded.
(b) The SDR is not applied to the LTEV_Loans but rather is applied to other figures arrived at in their calculation in accordance with “net present value methodology” [S.76(2)(d) of the Act].

Obviously I have had to make assumptions bcause the calculations have not been published as they should be.

Feel free to correct me.

By the way, a difference could arise between aggregate LTEV of properties and the Consideration paid where the poor quality of security led to a decrease in the value of the loans relative to the value of the property.

Thanks MOL

I will correct myself though:
“This is 0.85% outside the 5.25% expected.”
should read
“This is 0.37% outside the 5.25% expected.”

LTEV_All_NAMA_Loans should not be aggregated with All_Enf_DD_Costs when checking the application of SDR because the 29.5% calculation is an assumed pre-application of SDR LTEV. Therefore the Enf_DD_Costs should not be put back in when calculating the percentage.

BTW MOL – it never stopped you correcting me before (especially on long division)! 🙂

@Zhou

Is there an excel sheet that shows how all this turns €16bn into € 10.5 bn and then into €8.5bn?

Otherwise its a bit theoretical, no?

@L_R

There can be no Excel Sheet because the Govt has not seen fit to publish the mathematical models used to calculate
– LTEV_Indiv_Prop from CMV_Indiv_Prop
– LTEV_Indiv_Loan from LTEV_Indiv_Prop

Therefore we can only speculate and interpolate. However, we can deduce what factors might be applied during different stages of the calculations by reference to the legislation and regulations.

We can further deduce general relationships from the figures furnished. For example, we know that the enforcement and due diligence costs have been frontloaded onto the first tranche which represents approximately 20% of all NAMA-bound loans.
We also know that the consideration paid for the loans is well below the long term economic value of the underlying property.
However, that gap (between the LTEV of the properties and the actual consideration paid) does not account for all of the enforcement and due diligence costs.
Accordingly, one can deduce (subject to the information provided being correct) that the long term economic value of the loans in the first tranche must be greater than the long term economic value of the underlying properties.
This in turn shows that the safe guards and limits on the long term economic value of properties offer more limited protection to the tax-payer than was previously thought.

The fuirther benefit is that we start to see where the weaknesses in the valuation methodology are. This allows us to ask the correct questions in order to try to achieve effective oversight of the process. It also allows us to point out that there is not currently effective oversight and that the process is something of a black-box.

One last comment before I go:

T1 includes 20% of Total NAMA loans.
Accordingly, approximately 20% of Total Enf_DD_Costs should be factored into LTEV_Loans for T1.
This would mean that approx €2.23bn Enf_DD_Costs (80% of €2.788bn) are the amount frontloaded from the rest of the loans into T1 discounts.
The difference between LTEV_T1_Prop and Total Consideration is €2.0bn.

Does this mean that LTEV_T1_Prop approximately equals LTEV_T1_Loans?

Or does my premise that LTEV_Loans > LTEV_Prop stand up on account of the €0.2bn discrepancy in this scenario?

What’s €200 mill between friends?

Considering that the cost of funds has to be applied in moving from LTEV_Prop to LTEV_Loan it suggests to me that the €200 mill is significant as the cost of funds discount must come on top of the enforcement and due diligence costs. Accordingly, I still think LTEV_Loan > LTEV_Prop.

Will Brian Lucey and Kevin Denny really agree to me getting an original copy of the signed article? I have my doubts…

The distribution of the first tranche is skewed towards investment properties- 65% investment and 9% hotels. In the original NAMA BP, land and development was 65% of the total. Moreover only 58% in T1 was in Ireland compared to 65% in the BP. So far the worst performing asset class seems to be under represented in the sample.

More possible calculations…

IF we assume that
(T1_Consideration + (Total_Enf_DD_Costs – 5.25% of LTEV_T1_Prop) = LTEV_T1_Loans

THEN (all figures in €bn)

Total Loans €81.0000
5.25% Enf €2.7880

LTEV Total Loans if Enf reps 5.25% of same €53.1048
Total Haircut €34.4386

Loans T1 €16.0300
CMV Prop T1 €9.4400
LTEV Prop T1 €10.5100
Consid T1 €8.5100
Approx Enf T1 (5.25% LTEV Prop) €0.5518

Balance Enf €2.2362

LTEV Loans (Consid T1 + Bal Enf)* €10.7462
* Note – this is estimated LTEV loans after the application of SDR and CFDR

Approx Haircut T1 32.96%
LTEV_T1_Loans (Consid T1 + Bal Enf) €10.7462
Approx Haircut T1 32.96%
LTEV_T1_Loans: CMV_T1_Prop 113.84%
LTEV_T1_Loans: LTEV_T1_Prop 102.25%

@ Zhou

Could you explain what the “NAMA – upfront recovery of costs on
total portfolio” are?

What are the €2,655m of Recovery of enforcement costs?

Many thanks.

@L_R

See the second page of the following document:
http://www.nama.ie/Publications/2010/NAMATranche1.pdf

The EU Commission insisted that the Haircut include a 5.25% discount to provide for enforcement costs and due diligence. This was previously included but at a lower rate. One would expect this to be levied on the LTEV of the loans. In any event, it appears that the initial fugures have applied this discount for all NAMA loans to the first tranche.

If one were cynical one might speculate that the only difference between the haircuts forecast in Sept 2009 and those announced yesterday is this increased discount. In that case, it is hard to know why the Minister said things are much worse than expected and why the banks are kicking up. Is it all just choreograpy?

@Zhou_EnLai

Yes – this has been very carefully thought out in political terms in terms of “choreograpy” … the timing is too neat (i) the few questionings in full TV glare – and the bought GP poor fools (ii) the move on Quinn (iii) the sheparding in the Dail (awesome on Oireachtas live – looked like they were doing the sheeply shopping) (iv) the timing of BoI and especially Anglo Results (v) the Easter break and the benefits of time (vii) and so on. Time, and dragging out time, appears central to political strategy ….. and one assumes loads of dangerous evidence into the black hole and safe hidey_holes for dodgy loot.

Good Sums above – but you already hold one award for March (-;

It is a shame.

A shame.

All this pointless mental diddling when an ounce of prevention could have prevented it all.
Does anyone appreciate the nature of banking yet? The “business” cycle?

What hangs on this?

Fiddling while Dublin burns.

I know its April Fools, but is it really the case that the much vaunted larger haircut of 47% is misleading in that the first tranche of loans have had entire enforcement and admin costs deducted from them? If so, is Zhou the only one to have spotted this? What about the media, or indeed Moodys?

@gadge

See Pat Macardle spinning in today’s IT that the rest of NAMA will also get hair-cut of 47% – logic, and limited data, suggests the hair-cut much tighter and essentially ‘bald’ in a good few ‘land’ cases.

@ D O’D

Interestingly Pat McArdle makes reference to 5.25 per cent being deducted “upfront”, but does not make it clear whether he understands this to mean that the 5.25% of the entire of the loans to be transferred (approximately 81 bn of loans) has been deducted from the payment for the first tranche (approx 16 bn of loans). If the enforcement and admin costs of the entire of the loans has been deducted, then it would seem to follow that the haircut of 47% on these loans is likely to be higher than for the balance of the loans: this would seem to follow from the fact that the LTEV of the balance of the loans will not have to be reduced by reference to 5.25%, their “share” of the enforcement and admin costs having already been deducted upfront at time of the transfer of the first tranche.

The other point about Pat Mc Ardle’s article is his assertion that senior bonds are “sacrosanct” and rank equally with deposits in a winding up. Am I not correct in thinking however than prior to the early part of September 2008, the maximum legal protection for deposits was 30k. This was extended to 100k, and then ultimately guaranteed in full on 30 September 2008. If this is correct then, prior to September 2008, bondholders could not have been in a better position than deposit holders and had no meaningful legal protection. Of course, this changed on September 2008, and indeed it may be argued that there was an implicit guarantee. However, given the limited legal protection afforded to even deposit holders prior to September 2008, it does seem to be stretching the point to say that senior debt was “sacrosanct” .

@gadge

Simply observe who speaks in the interests of the well-connected vested interests – obfuscate, deflect etc and hide the evidence is their only game in town – and they are winning – hence the need to continuously challenge …..

It is intersting to refer back to the “Supplementary Documentation” of 16 September 2009. At that stage NAMA estimated an average 11.7% interest roll-up. It also estimated a average 47% decline in property values (in Ireland and abroad).

Using the Tranche 1 figures, and assuming 12% interest roll-up and an original LTV of 77%, one comes up with an approximate decline in property values of 49%.

Note that in September 2009 a circa 30% haircut was predicted for a 47% decline in property values! Also, given that more interest has rolled up we are probably closer to the 47% decline.

If these calculations are right (i.e. if the loan balances include rolled up interest) It is hard to see where the big shock was that makes things so much worse!

The calculations are incorrect if the “book value” in the Tranche 1 information does not include rolled up interest. “Book value” did not include rolled-up interest in the Supplementary Data. However the detailed Tranche 1 information refers to “loan balances” rather than book value. Accordingly, I think rolled-up interest is probably included. I would be grateful for other views on this point.

What if the “loan balances”/”book value” in the First Tranche Data do not include rolled up interest? Then the average decline in property values is 54.7%.
Irish loans only made up 58% of the loans transferred.
It was estimated in Sept 2009 that Irish Property had declined by 50% and that Irish loans made up 67% of the land and development loans and 67% of the associated loans.
From the average decline of 47% we can interpolate that foreign underlying assets must have declined by approx 41.9% [((67*50%)+(33*40.9%))/100 = 47%].
If we assume that the 40.9% decline in the value of underlying foreign assets has held steady then the average decline in Irish underlying assets must be 64.7%. [((40.9%*42)+(64.7%*58))/100=54.7%]

I think it unlikely that NAMA is estimating a 64.7% decline in Irish property values as of Nov 2009.

Therefore, I conclude that the loan balances include rolled up interest, that the decline in property values is as expected and that the haircut is as expected save to the extent that the standard discount rate and cost of funds has increased.

[PROVISO:
The only additional factor that could affect the consideration paid appears to be the quality of security. I don’t think security can really be as bad as they say. A little transparency on this issue would help. In particular, one wonders how the discount is affected by the security. Is there a formula? The valuers would have no role in that.

In the meantime, I am very sceptical about the First Tranche haircut really reflecting a much more strinent haircut than expected that will apply to future tranches.]

Comments are closed.