Irish Economy

Do the NAMA figures add up? A broader and more realistic assessment of long-term economic value

17 Sep 2009

"NAMA's entire estimation of long-term value hinges on this subset (Dublin) of a subset (commercial projects) of a subset (projects underway) of a subset (Ireland) of the total loan book."

Earlier today, I posted some basic facts about NAMA. This post builds on that by trying to work out where NAMA gets its €47bn current market value and more importantly its €54bn ‘plateau’ value, through looking at the numbers in detail. I’ve started this post not knowing how it will end, so I’ve no prior beliefs on this. I’m trying – and no harm reminding everyone that what I write here is as an ordinary punter with a degree in economics and not in any sort of official capacity as economist – to put to best use the numbers made public yesterday, so if anyone spots any errors, let me know and I’ll be happy to correct them. OK, let’s go!

Who lent what where?

The pie chart below shows the regional breakdown of the €68bn lent out. As you can see, Irish loans account for two thirds of the total land and developments that NAMA will control. This gives us one important guidance for further analysis, namely that discussions that focus on local market conditions in Ireland are only relevant for two thirds – or €51bn – of the total loan book value. (Arguably, the fate of the Northern property market is very similar to the rest of the island, so one could make the case that we could look at €56bn.) In terms of who lent what, Anglo (now in State ownership) accounts for 30% of total lending, while AIB and Bank of Ireland account for a further 60% between them.

NAMA land and development, by bank

NAMA land and development, by bank

What is the breakdown of Irish loans?

The detailed figures published yesterday give a total of €37.5bn of Irish land and development assets. Whatever about confident assumptions about the US, UK or European property markets, what can we say about the future for these Irish assets? This is singularly important given that the NAMA document rests crucially on just one fact: yields on commercial property in Dublin are above their long-term average. Is that good enough?

First, let’s look at what’s out there in grass and fields, as opposed to bricks and mortar. A breakdown contained in the Department of Finance document published yesterday shows that the split between land and development is 56%-44%. (The full breakdown is: land (36%), development (28%) and associated loans (36%), but presumably the latter is just a derivative form of the first two. I work this back in, in a pro rata way, later, but if anyone knows better, just shout.)

This means that, before we even break the loans down to commercial versus residential, we have €21bn, or more than half accounted for, by land. This is the very same land that we have heard has been devalued by anything up to 95% from the peak, due to the probability of a good deal of land being rezoned to agricultural. It’s reasonable to assume that this land splits down in some way between that which probably will be developed (and which has presumably fallen in value by 50%), and that which probably won’t (likely down 90%).

This leaves a further €16.4bn of loans which have as their collateral construction development projects already underway in Ireland. This is presumably the chunk of NAMA that led to the inclusion of some yield analysis, as mentioned above. However, this figure itself needs to be broken down into residential and commercial. The detailed figures at the back of the draft legislation do not give this breakdown for any bank apart from Anglo Irish Bank. Its loan book is split almost half and half between the two (to be exact, 51% residential, 49% commercial). Should this ratio prove to be typical for other banks also, this means that about €8.4bn of the loan book is in Irish residential construction projects at least partially underway, while €8bn is in commercial projects underway.

Amazingly, it is on this latter commercial segment of the market, worth from what I can see about €8bn – and only this €8bn alone – that is the subject of any form of yield analysis. The NAMA document looks at yields on three sub-segments within commercial again, retail, office and industrial property, in one location (Dublin). It compares these yields with a range of other cities in Europe and with the long-term Irish average, finding that yields in August 2009 were slightly higher than both.

But it must be remembered that NAMA’s entire estimation of long-term value hinges on rents holding constant in this subset (Dublin) of a subset (commercial projects) of a subset (projects underway) of a subset (Ireland) of the total loan book. If Dublin was soaking up half of all commercial projects, this would suggest that NAMA is basing its entire long-term economic value on perhaps 5% of its proposed loan book.

What about long-term economic value?

Let’s see if I can recap and re-estimate long-term economic value. To keep things simple, we can say there are four types of loan that would go on NAMA’s books (the numbers in brackets are my best estimate of peak value, done by attributing associated loans, so the numbers add up to the €88bn original value reported):

  1. Irish commercial (€11.8bn)
  2. Irish residential (€12.4bn)
  3. Irish land (€31bn)
  4. All foreign (€31.6bn)

The two questions we need to know for each, to start discussing long-term economic value, are (1) how far have property values fallen so far, and (2) how far out is the yield based on current prices?

Let’s assume that NAMA and Jones Lang LaSalle got it right for the €8bn in commercial developments, i.e. that prices have fallen by 47% on average so far and that they will rise a little as yields level off to historical and European norms. This means that the €11bn in commercial projects in Ireland is now worth €6.2bn but will rise to €7.5bn.

What about the €12.4bn in residential property? There are no figures that suggest that the average fall anywhere, let alone the country as a whole, has been 47% yet. Based on discounts from asking price data, 35% is a more accurate average of the fall from the peak. However, the average yield in residential property is just over 3%, only about half a typical yield of 6% (this yield is by no means overly generous).  This means that while the properties are probably worth €8bn now, their adjustment to long-term economic value would mean a downward adjustment to €4.2bn.

Regarding land and overseas projects, the value of Irish land has fallen dramatically and in many cases will not increase significantly as planned developments are scrapped. Overseas, property markets may be turning, having proven more flexible (in the case of the US) or less inflated (in the case of much of Europe). Let’s put to one side, then, the majority of NAMA’s loan book, i.e. Irish land and foreign properties of all sorts, on the assumption that they already reached approximately their long-term economic value. This means that based on what’s happening in residential and commercial property, one would expect a further 5% fall in the value of NAMA’s collateral – and not a 15% rise – as the market proceeds to its long-term economic value.

One further complication is rents, however. Rents in both commercial and residential segments of the market are falling, due to oversupply. If rents were to fall, say, 20% over the coming three years as they find their new level, this would mean a further fall in the value of commercial and residential property on NAMA’s books of €2.5bn.

In short, even if you think values for Irish land and foreign property have levelled off, estimating the true value of the €24bn in commercial and residential developments in Ireland does not translate into a straightforward mark-up of 15%. In fact, if you believe (a) that 6% is a healthy average yield for Irish property (as NAMA itself does), and (b) that it’s likely rents will fall a further 20% as the market handles recent overconstruction, the adjustment from current market value  should be downwards by 10% to about €44bn, and not upwards to €54bn.

Does anyone know if there’s a 1% commission on saving the State €10bn?!

Tags: , ,


  1. Graham Stull said on September 17, 2009 | Permalink


    Excellent stuff.

    As you demonstrate above, there is no problem with using NPV rules to induce LTEV values for Irish property – the problem is the political fudge.

    Ideally, we should start by asking 1) what level of residential and commercial rents does Ireland need to descend to in order to restore competitiveness? Then 2) based on (1), what is the value we can back out for the underlying securities on the Nama loan book, using an NPV rule on a 6% yield.

    (Voice from the tent): “Ah, but you see, there’s demographics…and using 1998 as a reference point…and factoring in the luck of the Irish…”

  2. John O'Connor said on September 17, 2009 | Permalink

    Absolutely delighted to see someone tackle the figures like this, thank you.

    Small point regarding your assessment, I don’t think it’s prudent or neccessary to pro-rate the associated loans between land and development. Prudent, because we just don’t know; not neccessary, as I think the strong message of devalued unproductive land asset values comes across with just 36% or 50+%.

    If the government want buy-in from an informed public for the LTEV argument, there topline 10% argument is not enough. They need to complete the analysis split by region, by asset class, and by productive capacity of the asset.

    As if a lot are unproductive assets, we will need a lot more than a 10% increase over 10 years to cover the interest bill alone.

  3. Ronan Lyons said on September 17, 2009 | Permalink

    Hi John,
    Thanks for that – agreed, my aim was to not confuse the issue by introducing different totals of €54bn and so on.
    As discussed on, the good news is that you can take out the associated loans, and the headlines and percentages stay the same, the only thing that’s different is the size of the pie.

  4. adrem said on September 17, 2009 | Permalink

    According to DAFT (do you still work for them) the average rent on a 1 bed apartment in Dublin 1 (just as an example) is c900 per month. According to you that represents a 3% yield – implying a purchase price of 360,000. However in actual fact, again according to DAFT I can purchase a 1 bed apartment I can buy such property for c175k. That would imply a yield of 6%.

    What am I missing here Ronan – your own numbers as at mid August as published on Daft seem to contradict completely what you have posted here today? I did the same check against 3 beds in D24 – again the yield using DAFT average rents and current prices is more than 6%.

  5. Ronan Lyons said on September 17, 2009 | Permalink

    Hi Adrem,
    Thanks for raising this – one quick thing I should note here is that what I’m writing on this site, which is my personal site, is an economist’s thoughts on the world around him. It shouldn’t be taken to be an official daft position or anything like that.
    All that said, one thing I would direct you to is the yield series, published each quarter towards the back of each rental report. It calculates the yield by bedroom segment and region of the country, just what you’re looking for I think. You’ll find that, as of Q2 2009, the highest yield in the country is in Dublin city centre (4.6%), while West Dublin is 4.2%. The national average yield is just 3.3%, though, down from 3.4% earlier in the year because of the speed at which rents are falling.

    The important thing is not to find one example that makes sense, it’s to find the average and see if that makes sense.

  6. Henry Withinshaw said on September 17, 2009 | Permalink


    I agree with you and think that you have highlighted something that is very important in the valuation debate. I have written about residential property yields a couple of times on but only from an anecdotal perspective. I also feel that there is a significant way to go.

    However, I do wonder if you are not underestimating the current fall in property values. Your 35% reduction from peak correctly reflects asking prices, but asking prices do not reflect market value (though they may be the best we have to go on, I admit).

    Although reluctant to do so, I would expect honest estate agents in Dublin would confirm that realised sales currently would suggest a 50% discount to peak – though I realise that this doesn’t reflect a country average.

    Although I think a 6% yield would be optmistic, even a 4.33% long term normal expected yield would bring a similar answer to the one you are getting at €4.2bn. 6% would be considerably worse (€3.1bn).

    But when we’re only talking a difference of €1.1bn, who’s counting??

  7. Adamirer said on September 17, 2009 | Permalink

    Excellent work. The 47bn is never an accurate figure to begin with, I just hope some of this level of analysis seeps through to the national media.

  8. Ronan Lyons said on September 17, 2009 | Permalink

    Henry, thanks for the comment.
    I rounded significantly up from the average fall in asking prices nationwide (while Dublin is indeed falling fast, Munster is playing hard to get!), which is about 25%, i.e. is it reasonable to assume the typical buyer gets somewhere in the region of a 10% discount from the asking price at the moment?

    The other thing I would suggest, as you yourself note, is that Dublin-based estate agents may well be right in talking about falls of 40% or more, but I would like to hear more from country estate agents, who are discounting much less and whose clients are seeing a much larger time to sell.

    @Adamirer, thanks for the comment.

  9. Henry Withinshaw said on September 17, 2009 | Permalink


    Many thanks for this explanation. It helps a lot in my undertstanding of the national picture.

    A couple of thoughts for the sake ot it! 1) I think your 10% is fair. 2) It would seem that country estate agents are not discounting enough (out of fear, and hope, I suspect). 3) My instinct in South Co. Dublin is that, on the whole, many areas are woefully over-valued still, but I expect the recived opinion of significant (stupid?)premium for better areas will soon be a thing of the past. I know it’s just one area but I would expect it to have a fair few development loans for NAMA in it. Now that would be an interesting map!!

  10. david mc williams said on September 17, 2009 | Permalink

    excellent work ronan, i don’t know whether to laugh or cry! all the best david

  11. Fergus O'Rourke said on September 18, 2009 | Permalink

    Super stuff, Ronan

  12. Stephen.kinsella@gma said on September 18, 2009 | Permalink

    Excellent article Ronan.

  13. Damien said on September 18, 2009 | Permalink

    Congrats Ronan, Excellent work. What really scares me is that it appears that no one in the Dept of Finance, NTMA etc. is listening (or has a calculator!). The anti-NAMA arguments put forward by you, David McWilliams, Brian Lucey, Constanin Gurdgiev, Dermot Desmond and many others have not been rebutted yet NAMA keeps rolling forward.

  14. kevin denny said on September 19, 2009 | Permalink

    Ronan, I wouldn’t bet on getting your commission- nice try ‘though. Even a .001% penalty for wasting €10b might change some people’s behaviour.

  15. adrem said on September 19, 2009 | Permalink

    Ronan I have to disagree !!! If you take your DAFT report and gross up the average rent by the yield also shown in the DAFT report you get prices way above the current sale prices of properties again shown on DAFT. In other words the prices have fallen by more than you are allowing for and thus the current yields (all based on a single source of data – DAFT rental levels and DAFT house prices) are higher than you are showing above.

  16. Ronan Lyons said on September 19, 2009 | Permalink

    There’s not a lot more I can do for you, as we’re both using the same data and getting different results! All I can do is restate is that looking over the entire sample of daft properties, the average yield in the country, using Census 2006 weights, is 3.3%, which is much closer to the ‘just above 3%’ I quoted than it is to your 6%.
    One other comment I’d make is that both are based on asking price, so if you’re prepared to make discounts from the ask for house prices, you have to do the same for rents, although the size of the discount of course may be different.

    One could argue that yields in the more important areas for the lettings market, e.g. Dublin city centre, are higher. That’s certainly true, but the lettings and sales markets throughout the country have become much more integrated over the past 5 years – i.e. you can buy or rent in an segment of the market. Therefore I think it’s reasonable to say that the Census-weighted average yield is a very pertinent measure.

  17. adrem said on September 20, 2009 | Permalink

    Ronan – that’s not a reasoned response – you are supposed to be using data to rationally make an argument – the data simply doesn’t add up though. I’m not discounting anythin – I’m simply picking rents from Daft and house prices from daft dividing one by the other and you get a yield that is much closer to 6% than 3%. Equally if you gross up the rents by the 3% yield you are using you end up with prices massively over the factual current market prices. These are facts Ronan – easily verifiable.

  18. Ronan Lyons said on September 20, 2009 | Permalink

    To be fair adrem, I was being quite generous in my last response. In truth, you are using a handful of samples, perhaps no more than 5, to reach your conclusions.

    On the other hand, the method employed in calculating the yields is a series of econometric regressions, controlling for measurable attributes of the properties in question and – as is done by the ESRI/PTSB – is two-stage to exclude outliers. The regressions’ estimates are all highly statistically significant. The most telling part, though, is that the sample sizes are tens of thousands times the size of yours. For example, the 2nd quarter of 2009 comprised just over 99,000 observations across sales and lettings.

    In sample sizes that large, of course you are going to be able to find at least one observation that matches what is clearly your prior belief about yields for residential property. My job, however, is not to validate anyone’s prior beliefs but to uncover as best as possible the true figures. So for every matched pair you find that has a yield of about 6%, there is some other property out there with a yield of say 2%.

  19. adrem said on September 21, 2009 | Permalink

    Ronan, I’m not going to continue arguing with you on it – your numbers are wrong – fact. Not my estimation – a fact. Look at the average prices for any of the regions or house prices on Daft or on Myhome and compare that to the rentals shown on Daft divided by 0.03 and you get significantly higher figures using your 3% yield than the factual average house/apartment prices. I agree 6% is probably too high but the real number is closer to 4.5% than to 3%. And whilst in nominal terms that’s a small difference – as an economist you know that a 50% higher yield is a fundamental difference.

  20. Ronan Lyons said on September 21, 2009 | Permalink

    Agreed, as I noted a few comments ago, this is clearly going nowhere, as you have your prior beliefs and no amount of evidence is going to shift you. I am happy to send on my regression do-files, so you can see the method employed and you’re more than welcome to make any constructive comments to improve it, but you can’t just continue make wild assertions with no evidence.

  21. adrem said on September 21, 2009 | Permalink

    Ah come off it Ronan – it’s nothing to do with prior beliefs – I’ve not made any wild assertions either, I’ve told you exactly what my calculation basis was – how much information do you want? The implied price (using your info and the DAFT rent) of a 1 bed in D1 is 356,800, in D2 it is 406,800 in D3 it is 340,000, in D4 it is 416,000 in D5 it is 326,800 in D6 it is 334,000 in D6W it is 328,000 in D7 it is 321,600 in D 8 it is 339,600 in D9 it is 325,200 in D10 it is 320,400 in D11 it is 331,600 in D12 it is 323,600 in D13 it is 343,600 in D14 it is 385,600 . . . . I’ll stop now – point being if you look to purchase 1 bed apartments in any of the above locations there are units available significantly below those implied prices. As you seriously trying to asset that the above are average prices in those locations? Come off it. Maybe the Daft rental data is not accurate but (again) I am not making these numbers up they come from your employer and from you, don’t attack the messenger.

  22. Ronan Lyons said on September 21, 2009 | Permalink

    One last time before I call troll:
    The last Daft rental report says:
    - page 7: typical rent for a 1-bed in Dublin city centre is probably about €950 (€900 in Dublin 1, €1,000 in Dublin 2)
    - page 11: the yield for 1-bed in Dublin city centre is 4.7%
    - so 11 times the rent (allowing for the standard month’s vacant/maintenance) = €10,450
    What average price gives you a yield of 4.7% on €10,450 a year? €222,000, about half the price you’re talking about.

    Now, do that for each region in the country and weight them by their populations and you get 3.3%.
    End of.

  23. adrem said on September 21, 2009 | Permalink


    I wasn’t trolling (and never have been) – you asked for comment and feedback and I gave it to you.

    I was using the same rent levels as you. At a 3% yield on 950 per month you get a purchase price of 380k.

    4.7% is significantly higher than the 3%.

    Also I wasn’t allowing for vacancies/maintenance in my numbers so I was applying the 3% yield to the full rental amount

  24. Ronan Lyons said on September 21, 2009 | Permalink

    I’m sorry to be harsh on you, but one of the first things I said to you was to look at the details of the rental yield and not just take the national average and find one example that seems to run counter to it. Look at the detail of the daft yield table, go back through past reports and see its evolution over time, look at how it differs across segments and you’ll see what you’ve done is either by accident or design pick a combination of the type (1-beds have highest yields in the country) and location (Dublin city centre has the highest yields in the country) that best portrays yields in a positive light.

    Unfortunately, the vast vast majority of properties in Ireland are not 1-beds in Dublin city centre.

  25. adrem said on September 21, 2009 | Permalink

    Ok ok – sheesh !!

    It wasn’t by design – I just went to the top of the table and it starts with Dublin so I went with that – also the highest population density and property densisty so I reckoned it was a useful proxy – obviously not so much !

  26. Baz Grant said on September 24, 2009 | Permalink

    NAMA in a nutshell – a short film:

5 Trackbacks

  1. [...] leading property number cruncher, Ronan Lyons, has a post essentially explaining how he would have done the LTEV calculations if he had been asked. Key [...]

  2. [...] showing potential flaws with the long-term economic value (I have two modest examples here and here) was zero. The estimate of long-term economic value is explained just as before: It is estimated, [...]

  3. [...] Smart Taxes Network member and leading property economist Ronan Lyons has asked if the numbers add up and has done his own calculations in relation to “long term economic value”. A summary of his conclusion is below and his excellent post can be read in its entirety here. [...]

  4. [...] just reallocates the flooding problem. The approximately €8.4bn in Irish residential projectscovered by NAMA translates into about 34,000 residential units, taking €250,000 as an average [...]

  5. [...] this is not 20-20 hindsight, this is using the information available at the time. Many people, myself included and on more than one occasion, tried to make NAMA aware of this at the [...]

Post a Comment

Your email is never published nor shared. Required fields are marked *





Ronan Lyons Posts RSS feed

Receive RSS updates of Posts and Comments.

Subscribe to Email Updates

Subscribe to Email updates.