The Macroeconomics of Long-Term Economic Value

Any chance of a reminder on what co-integration and error correction models are for those of us who didnt listen carefully enough to Dr Thom 20 years ago?

i wonder will the ecb b willing either 2 exchange hard cash for the second tier (the deferred paiment bit) and on what pricing basis or at least count (again on what basis) that promise bi the Irish govt/NAMA to pai the ‘ultimatel-realised loan value’ at some point in the future as part of the banks’ current capital. The 2 tier pament might act if done in a certain wai for instance as an incentive for NAMA 2 off load loans quickli (2 former developers perhaps or perhaps the likes of Frank Fahei could help NAMA out bi arranging purchase of some of Anglos loans in 4 example the US for instance). The advantage of this would b 2 get the loans off the NAMA balance sheet asap and mean NAMA could pai off the banks the second tier asap at a discounted (compared to offloading the loan in 20 ears) ‘ultimatel-realised loan value.

@ Phillip

Great piece as it opens up a new LTEV dimension. Back to what banks do. If there is an reasonable argument to pay more for a loan as the financed asset is required in the future. And if NAMA is to become a repository of needed assets then additional cost is justified – it becomes important then to identify those assets and only those assets. It also reclibrates NAMA’s mandate as its objective has two dimensions: the husbandry of needed assets and the elimination of others.

@Philip
As ever this is an elegant and well structured piece.

Two issues: first, as is commonly used, the models are linear. However, almost by definition the existence of a bubble implies nonlinearity or at least a regime shift in any underlying relationships. Indeed, you allude to this “Third, it is possible that the beta coefficients are subject to structural change: the estimated values over an historical time period may not provide a perfect guide to the future.” Indeed, the underlying parameters that drive the V* may well themselves be regime shifting. The consequence is that underlying (possibly stable) relationships may be hidden. As perhaps a guide to modelling then a non-parametric or stochastic cointegration approach would be required. These approaches can back out the adjustment coefficients over different times and regimes, which may well be needed when designing policies, assuming one can know or reasonably infer the regime one is in.

Second, I would be interested to hear your views on the design of the risk sharing mechanism. As I read it your note is a call for NAMA to incorporate same, which I would endorse. As put forward by (as we can now call him) Governor Honohan, on 30 August in the Sunday Tribune “In my scheme, the banks would get bonds only for what can be confidently expected as recoverable on the loans; in addition, bank shareholders would be given an equity stake in Nama’s future recoveries.” However, the key word here is shareholders. The DoF has signalled that they are looking at “variants” (another word for mutants….) of the Governors scheme. Two major parameters define the extent of mutation possible, on whom is the risksharing burden to be placed and in what form. In the first case there is a choice between the banks themselves as corporate entities and the providors of capital – Governor Honohan’s idea is for the latter to be used, but it is not at all clear that this will be the case. Were banks, as entities, to be given a stake in NAMA to the exclusion of bank capital providers, which would be a very strange mutation of the Governors scheme. Second, the form of the stake in NAMA’s upside given in the second stage is important. One can imagine a range of instruments in NAMA, from plain equity through hybrid instruments to debt claims of a different type to the initial payment. Again, these will give rise to very different incentive sets, and these differentials will themselves change as to whom they are given.

A little bit off topic.

I am beginning to get worried about the effect of

(i) using long term economic value and

(ii) trying to recover as much money as possible for acquired assets

on the liquidity of the market.

How can NAMA elect to dispose of assets at a loss which it has paid for a premium for?

Even if a NAMA 2.0 process is used, won’t NAMA still be obliged by the goals of the legislation and by its duties to former bank shareholders to act as if NAMA had a fiduciary duty to owners of those assets?

I have generally been worried as to how we will incentivise outsourced bankers to work out loans as profitably as possible. I am now becoming worried that if they are too prudent or too effective that the market will remain zombified.

I have some experience of constructing models to find the balance point for refrigeration systems. The science is well understood, we know that in reality the systems remain stable (mostly) but it is still a challenge and involves some sleight of hand sometimes to dampen it all down.

We all have experience of weather modelling and while it has improved to the extent that a 5 day forecast now has the accuracy of a 1 day forecast years ago, we know that advances wont be on the scale of Moore’s Law. And this for a system that remains fairly stable and where we can make broad brush forecasts by looking at the calendar.

Modelling long term economic value is certainly a worthwhile exercise.
We just need to understand the uncertainty in the system. Hurricane forecasts typically show alternative predicted paths using slightly different models to give an ever expanding cone of uncertainty out to 5 days.

We are faced with a system that is driven by human variables, that is poorly understood, and where for a variety of reasons we are in uncharted waters.

The initial indications are that the internationally traded private sector is responding quickly to the challenges, even the learned professions are inching towards economic reality but we will be waiting a long term for the public sector charges to drop – one thing we can be certain that public sector charges will not overshoot on the way down. These highly politicized decisions will have a very big bearing on how quickly the Irish economy can return to it new (much lower) long term potential growth.

We live in a small open economy that managed to divorce itself from the outside world for a decade as self-belief fuelled a bubble focused on the domestic economy.

Demographic factors; we had an enormous and reasonably well documented influx of workers from the accession states that has added about 10% to the workforce in the course of a few years. Will these people all stay? Will their friends and relations follow them? Will a generation of young Irish people leave the country? From memory, roughly 80% of the people born in Ireland in the ’30s or 40s emigrated.

Finally the scale of the bubble (I was worried in a general sense about property prices and over-inflated expectations from 2000 – the year of the 00 car) must pose the question; are we Japan or are we Sweden or do we have a lot of tulip bulbs in that wonderful new economic area, the Upper Shannon Basin.

We hear from NAMA that they are taking 300 variables from each loan.
It is all meant to sound wonderfully scientific.
LTEV will be determined by a formula.
These people may all be wonderful, but they face an impossible task.
The LTEV within any reasonable margin of error (10% even 20%) is under present conditions unknowable.

We live in a small open economy that managed to divorce itself from the outside world for a decade as self-belief fuelled a bubble focused on the domestic economy.

So, as an engineer I think you are right to try and model long term economic value. We do our best all the time with approximate models and we know we are not always right.
As a taxpayer, I don’t think we should be relying on modelling LTEV to create the biggest property company in the world.

The answer is to avoid the valuation problem.
The means either temporary nationalisation (LP) or good bank/bad bank(FG).

@Zhou, I would have thought that the cost at which capital is made available to NAMA should be set at a sufficiently high level so that profit maximising (or loss minimising) behaviour would ensure that it disposes of assets at a reasonable pace. I’m afraid I don’t understand the funding model envisaged well enough to know whether the actual cost of capital will be identical to the reported 1.5% coupon on NAMA bonds, but if it is it seems a bit low for this purpose.

@Con, Zhou
If NAMA is funded entirely by 1.5% coupon bonds (or iou’s as Patsy McArdle calls them…) that would be the cost of capital. As to whether that is what it should be, another question. IMHO it shoud be higher

@Brian, my reasons for uncertainty were two.
1) I’m not absolutely clear that the 1.5% is fixed.
2) I presume that a bond with a 1.5% coupon would be worth significantly less than par under most circumstances, so that NAMA’s actual capitalisation will be less than the par value of the bonds issued. If the cost of capital was calculated on the basis of the market value of NAMA bonds, rather than their par value, then it would presumably be higher than 1.5%. I know that the bonds will be deposited with the ECB as collateral, rather than traded, but I would have imagined that the level of funds advanced to NAMA by the ECB would shadow the market-equivalent value of the collateral, making the fact that they will not be traded moot.

It’s a long time since I studied any finance, so all lessons will be well received.

@Con
Indeed we dont know but we are well informed that the bonds will be at 1.5% coupon (for how long and how floating is still a mystery). Indeed, that should also be worth c 70% for a ten year bond compared to the 5% yileding bond that is the present state. So one of two things will happen
a) ECB will overpay , paying face for something that is worth 70% face
OR
b) they will pay 70% face in which case the state will have to issue 1.4 times the face to get the banks 1.
It would seem illogical to pay more than the true value of nama bonds but given where they will be coming from ….

@BL

Thanks for the clarifications. Could it work as follows:

1. NAMA pays to BoI €30bn bonds at 1.5% for assets valued at €30bn (LTEV).
2. BoI presents €30bn bonds (1.5%) to ECB. ECB gives €21 bn cash in return.
3. BoI can say is has €30bn capital thereby reducing the banks capital requirements notwithstanding that the ECB will only give €21 bn cash.

??

@Philip,

Your note sheds valuable light on these complex issues and banishes some of the polemics this topic attracts. Your separation of the macro and micro aspects of the LTEV exercise obviously makes sense, but where I retain some unease is at the micro level where the valuers will look at the deviation from LTEV of individual loans and the associated collateral. They will need to have the macro-derived LTEV disaggregated to the individual level.

I suspect the public concern about the NAMA valuation process is that some loans fund developments that would not make sense with a population twice as large and GDP three times as large and that a hopeful non-zero valuation will be put on these.

Btw, your note attacted the approval of Alan Ahearne in a NAMA-defence piece in today’s IT. I’m surprised Karl W has not posted already.

Banks are best able to do what banks do.
Why involve a transfer to someone else? To steal. Who is stealing is up to investigators. Who is the victim? Only the taxpayer. So it is not theft?

Let BOI take over AIB. Get the government out of this and make it do what it should be doing!

@Philip

I agree with your conclusion that the uncertainty surrounding long-term valuation of Irish property is a key argument against paying too much upfront now and perhaps giving shareholders a stake in NAMA’s profits, should there be any.

Two questions.

First, just to clarify. Am I correct that when you say “further reinforcing” the case for a two-part payment that this means you are not in favour of the current draft NAMA legislation on the grounds that it exposes the taxpayer to too much risk? Alan Ahearne has cited your paper approvingly on the grounds that it shows that LTEV “has solid underpinnings”. I pretty sure many casual readers will have interpreted this comment as implying that you support the govenment’s current proposals.

Second, even forgetting the difficulties in estimating LTEV, I’m not sure I understand the “prima facie case to employ long-term economic value in
determining the price for loans that will be transferred from troubled banks to NAMA”

I agree that sometimes markets over or undervalue assets relative to how they would be priced in the theoretical world where everyone is rational and fully informed. However, even if Irish property assets were currently undervalued (a dubious proposition) it is unclear why the government should decide that “fairness” dictates paying more than currently could be realised. Other sellers of assets do not get this benevolent treatment from purchasers so why should bank shareholders?

@Brian, Con, Zhou

The bonds will be issued at Par with a low 1.5% coupon reflecting the short duration – they’ll be issued at the front-end of the curve (like Bills) with e.g. 6-month maturities.

This will create massive rollover interest-rate risk based on the steepness of the current yield curve. In other words, by 2012 they’ll be paying a coupon closer to 3.5%.

6-month German BOBL’s are presently yielding 0.39%, so in essence the 1.5% coupon implies a sovereign credit spread of 111bps – which is about right (the Merrill-Lynch Sovereign Ireland Bond Index as of Sep 4 was trading at +116bps over Germany).

As the ECB is facilitating the funding of NAMA in this respect, it goes to show to what extreme lengths they are prepared to go to, to prevent a Sovereign default of an EA member (which would bring down the euro currency bloc as acolytes of Soros et al would then target Greece, Italy as well as the Euro itself).

@Derek, thank you, very interesting. I guess that means we should expect NAMA’s cost of capital over its lifetime to be 2/3/4 times the initial 1.5% coupon.

That could have an interesting impact on the size of haircuts, if factored into calculations of LTEV, or on the size of the public subvention if the banks do not have to eat it.

Karl Whelan “is there anyone out there who doesn’t agree with these facts” 4) “The ECB is not lending NAMA money at all as to do so would violate the EU Treaty’s prohibition of monetary financing of government.”

But it might be a ‘nod-and-a-wink’ job – you know indirectly banks staying as quoted companies and providing “monetary financing of the government deficit” QE by another name. Why not – desperate situtations requiring desperate (and sometimes outside the EU Treaty’s rules) measures.

Karl says “even forgetting the difficulties in estimating LTEV, I’m not sure I understand the “prima facie case to employ long-term economic value in
determining the price for loans that will be transferred from troubled banks to NAMA”
Estimating LTEV and intrinsic valuations should not be ignored simply because they are difficult to do. The problem is that the valuation methodologies used to value physical assets simply looks at market value (usually taken as some comparable ‘evidence’). This is what was the valuation profession does. And when this is inflated way above their true economic value we get bubbles. Any sensible attempt at economic value would have put the breaks on such madness (even though it is difficult). Financial assets can be valued using intrinsic free cash flow valuations based on projections of the economic drivers. Can’t see why this cannot be done with real estate.

18 replies on “The Macroeconomics of Long-Term Economic Value”