I’ve been puzzling over S&P announcement (and supplementary Q&A material)  the other day, lowering the stand-alone credit ratings of the major banks and the Banking Industry Country Risk Assessment (BICRA) score for New Zealand.  The BICRA was lowered from 3 to 4, on a 10 point scale, where the banking industries of Greece and the Ukraine score 10.

As I noted on Friday, the S&P ratings appeared somewhat inconsistent with the Reserve Bank’s 2014 stress test results, in which even some very severe adverse shocks did not generate loan provisions/losses that were large enough to induce material annual losses in any year of the scenario for any of the major banks.

As a reminder, S&P did not lower the actual issuer credit ratings of any of the major banks. Those issuer ratings incorporate the probability of parental support, and the possibility of government support, in the event that one of the banks got into difficulty. The issuer credit ratings remain at pretty respectable levels, with each of the big four banks at AA-.    This table, taken from sorted.org.nz, draws on Reserve Bank resources to summarise the ratings scale.

Capacity to make timely payment

Description	Standard & Poor’s scale	Moody’s scale	Fitch scale	Approx. probability of default over 5 years*
Extremely strong	AAA	Aaa	AAA	1 in 600
Very strong	AA	Aa	AA	1 in 300
Strong	A	A	A	1 in 150
Adequate	BBB	Baa	BBB	1 in 30

* The approximate, median likelihood that an investor will not receive repayment on a five-year  investment in time and in full based upon historical default rates published by each agency.

Source: Reserve Bank

So the chances of getting your money back, from the big New Zealand banks, are still rated extremely highly. That seems about right to me. The chances of the parent banks, and the Australian and New Zealand governments, allowing creditors of these major subsidiaries of big Australian banks to lose money seems very small.   In that sense, creditors of the New Zealand banks may be slightly better-placed than creditors of the Australian banks – there is no large parent, with concerns about contagion risks, standing behind the Australian banking groups themselves.

But the focus of last week’s announcement was not on the issuer ratings, but on the standalone ratings, and the wider environment in which the banking industry is operating. Standalone ratings look at the ability of the bank concerned to meet claims on it without extraordinary parental or government support. In other words, primarily, the quality of the loans on the bank’s books and the level of capital it has to absorb any losses.

And there, having read the S&P documents, and the limited media accounts of their follow-up comments, I am still puzzled.   Three of the big banks now have standalone ratings of BBB+ (while ASB is one notch stronger, and KIwibank and Rabobank are one notch weaker). As the table above suggests, a rating at that level is not much better than adequate. Based on historical experience it is really quite risky. When the Reserve Bank sets minimum capital requirements for banks – and all these banks have capital well in excess of the minimum – they are looking at something much more robust than that.

An S&P spokesman is quoted as saying ‘I think we are on the same page, more or less, as the RBNZ”. But I don’t see how they can be.    Here are the loan losses from the Reserve Bank’s stress test scenarios.

And here are the key capital indicators from the November FSR.

Capital ratios are much higher than they were in 2007 (and the risk weights are now typically more demanding as well), and there is a substantial buffer over the regulatory minima. The Reserve Bank’s stress test loan loss estimates could be understated by half, and the soundness of the banking system as a whole would still not appear likely to be in jeopardy.

But there are other puzzles in the S&P material:

There is no mention at all of the banks’ heavy dairy sector exposures, even though risk on that book is no longer a “low probability event” but something that is crystallising now, as weak world dairy prices and a continued high exchange rate combine to create severe cash-flow difficulties for many farmers, and the likelihood of a significant reduction in the value of the collateral banks have. After all, in a case of somewhat mixed messages, even the government is now belatedly talking of selling dairy farms.

If there is one area where one might be a little critical of the Reserve Bank’s stress tests it is around the dairy scenario, which assumes payouts of just over $5 for several years. Payouts at that level would be not much different, in real terms, from the longer-term historical average levels. There is a real risk already of something more severe than that scenario.   I think the Reserve Bank scenario came out this way because it was built on a severe recession (to trigger the big housing market correction and sharp rise in unemployment). A severe recession would be likely to see a much steeper fall in the exchange rate than we’ve seen so far, which would support the NZD value of returns to dairy farmers. In other words, potential losses on dairy debt aren’t additive to potential losses on housing loans – the scenario in which one might be very bad will typically be offset by less bad losses on the other portfolio. In the current climate, the dairy books look rather sick, while housing losses remain trivially low. If the housing situation worsened markedly (unemployment rose to anything like the 13 per cent in the Reserve Bank’s scenario), the exchange rate would probably be revisiting the lows seem in 2000 (around 50 on the TWI) and net returns to dairy farmers wouldn’t look anywhere near so bad.

But having said all that, it is still surprising that S&P didn’t even mention dairy exposures, by far the largest non-housing economic exposures on New Zealand bank balance sheets.

And S&P still regard any major rapid correction to house prices as “unlikely”.   From his public comments the Governor appears to think the risk is rather greater than that, even while reporting stress test results suggesting that the banks can cope with such a shock. It is all very well for S&P to talk about how a major house price correction would also have wider adverse economic ramifications. Everyone probably agrees with that – at least, unless the house price fall resulted from regulatory liberalisation, which might well be net stimulatory – but New Zealand authorities have far more room to lean against a severe economic slowdown than do authorities almost anywhere else in the advanced world (most of whom have interest rates stuck near zero already). Even relative to Australia (of which more below), the Reserve Bank has an additional 100 basis points available to cut the OCR.

And in either country, interest rates cut to zero would be expected to result in a very large further fall in the exchange rate. Puzzlingly, in their supplementary material, S&P highlight such a depreciation as a risk. They talk of a possible fall in the exchange rate, following a sharp house price fall, as “damaging confidence and potentially limiting monetary policy flexibility”, but they provide no basis for these interpretations. Since both banks and borrowers are pretty well-hedged to exchange rate fluctuations, a sharp fall in the exchange rate would almost certainly be a helpful mitigant. And I’m not aware of any floating exchange rate country in the 2008/09 cycle where the fall in the exchange rate was regarded as problematic or a constraint on monetary policy flexibility. Perhaps Iceland is an exception, but no one thinks of the New Zealand banking system as akin to Iceland’s (even today Iceland gets a BICRA score of 7 from S&P).

Some of the cross-country perspectives don’t make a lot of sense either.   Just a couple of weeks ago, S&P was talking of upgrading the standalone credit ratings of the Australian banks, from A to A+, if those banks raised more capital to meet the higher minimum risk weights APRA is to require. But, as the IMF has recognised and is I think well-understood by the banks themselves, risk weights on housing loans have been materially more demanding in New Zealand than in Australia (or other advanced countries), thanks to the appropriately conservative approach taken by our Reserve Bank. APRA’s latest announcements probably only have the effect of bringing risk weights on Australian mortgage loans up to towards those already in use in New Zealand (the AFR reports average risk weights will have to be increased from a minimum of 16 per cent at present to a minimum of 25 per cent in future). House price inflation has also been strong in Australia – Sydney prices are even more unaffordable than those in Auckland –  and, if anything, credit growth has been running faster in Australia than in New Zealand. It is difficult to see quite how, on the material they have put out, the New Zealand subsidiaries warrant standalone ratings that would be three notches lower than those of the parents.

Finally, S&P defended the BICRA score (the move from 3 to 4) by noting that only 23 other countries had higher scores than New Zealand. That doesn’t sound too bad – although a far cry from “one of the strongest banking systems in the world”) if one thinks of 200 countries in the world, but in fact they only seem to do BICRA scores for about 85 countries. According to this document, from last November, only Switzerland scores a 1. At the time, countries with a 3 were Chile, Denmark, France, Korea, Netherlands, New Zealand, the UK and the US. That was, perhaps, questionable enough company given that Denmark, Netherlands, the UK and the US had all had banking crises in the last few years, and France was saved from one by the Greek bailout in 2010.   In group 4, along with New Zealand now, we find the Czech Republic, Israel, Kuwait, Malaysia, Mexico, Oman, Peru, Qatar, Slovakia and Taiwan. By contrast, Australia (and Canada) still scores a 2. I struggle to see how the risks in New Zealand look more similar to those “4” countries than they do to those in Australia and Canada (or even the UK). In each of those latter countries, house prices – at least in major cities – have got detached from sensible unregulated longer-term fundamentals.   And even S&P reckon that New Zealand house prices are not a “bubble” about to burst. Each has a floating exchange rate – a major element in economic resilience – and New Zealand banks seem particularly strongly and conservatively capitalised (the combination of risk weights, and actual capital ratios). Historically, residential mortgage loan losses not played a key role in systemic bank failures, and yet they are far the biggest exposures on New Zealand bank balance sheets.

The S&P model appears to put quite a lot of weight on New Zealand’s relatively high negative NIIP position. But I think they are largely wrong on that score too. First, the NIIP/GDP ratio has been fluctuating around a stable average for 25 years now. That is very different from the explosive run-up in international debt in countries such as Spain and Greece prior to 2008/09. But also the debt is largely taken on by the government (issuing New Zealand dollar bonds) and the banks. No one seriously questions the strength of the government’s balance sheet, or servicing capacity, even after years of deficits. And the ability of banks to borrow abroad largely depends on the quality of their assets and the size of their capital buffers. If asset quality really is much poorer than most have recognised, rollover risk could become a real problem, but it isn’t really an independent source of vulnerability

I’m not sure where I come out on all this. Even though the Governor is acting as if he doesn’t really believe the stress tests, neither he nor staff have given us any good reason to think their estimates – suggesting a highly resilient banking system at present – are wrong. Then again, people pay substantial amounts of money for S&P’s ratings services. It would be more reassuring if, in the follow-up commentary S&P had directly and specifically explained why they found the stress test results unpersuasive, or even why they think the risks to New Zealand banks are so much greater than those in, say, Australia or Canada.  Banks just don’t fail when they are (a) privately-owned, (b)  operating in a fairly stable environment (ie not newly deregulated), (c) not subject to government pressures to take on inappropriate risks, (d) where total assets have been growing no faster than nominal GDP for years, (e) strongly (and increasingly) capitalised, and (e) where the largest component of assets is residential mortgage loans.

As some commenters have pointed out, having been somewhat burned in 2008, perhaps rating agencies have an incentive to overstate risk at present. But even if that is true – and I’m not sure it is, after all the last few years have been characterised by a renewed and rather desperate global search for yield, any yield – it still can’t explain why S&P seem to see so much risk in New Zealand relative to the situation in other floating exchange rate advanced countries. Of flexible exchange rate OECD countries, only Turkey, Iceland and Hungary now have BICRA scores worse than New Zealand.

I don’t believe it.

AUT University has, over the last few weeks, been running a series of short essays on housing-related issues, in their Briefing Papers series.  There are now seven contributions from a range of different perspectives –  from economics, to social housing and health issues.  My contribution to the series is up today.  In it I reprise (briefly) my story that persistently high house prices, especially in Auckland, are best seen as the result of policy blunders of successive governments: land use restrictions running head-on into high target rates of inwards non-citizen migration.

For regular readers there will be nothing new in the latest piece.  For others, a fuller version of that story is here, and a complementary piece explaining why Reserve Bank investor finance restrictions are not a sensible or appropriate response to a problem of this nature is here.

David Parker has an interesting piece in the Herald, on how the various free trade agreements New Zealand governments have signed affect the ability of New Zealand to restrict non-resident purchases, should it wish to do so.  The heart of his argument is here:

The most favoured nation provision in article 139 does apply to existing investments and controls on new investments. If we want to further restrict the sale of farmland or houses to Chinese investors, we can. Article 139 simply requires NZ to treat China no less favourably than other countries. Clause 3 of article 139 means earlier agreements with our Australian and Pacific Island neighbours are not affected, and do not flow into the China FTA. Later agreements do flow through.

National does not believe there should be more restrictions on foreign buyers, and so the South Korean FTA does not contain the protections found in the China FTA. This creates risks if New Zealand moves to restrict or ban South Korean investment in residential property. Screening or bans are allowed for existing categories but not new categories, that is farms but not houses.

Article 139 of the China FTA means NZ can’t properly ban sales to Chinese investors but allow them to South Korean investors. Even the South Korean FTA does not limit the sovereignty of a future New Zealand government to restrict house sales to foreigners, but it does create a risk of South Korean claims.

If Parker’s reading is correct, it does appear to leave some options open. According to the MFAT website, the New Zealand-Korea FTA has not yet been ratified, and so is not yet in force.

Rodney Jones proposed a 20 per cent stamp duty on non-resident purchases in Auckland. Such restrictions or taxes are not a first-best solution. Particularly if the non-resident demand from China is likely to persist over the medium to long term, it would be much better to liberalise land-use restrictions and make it much easier to supply new houses and apartments. It is an export industry.  (But if the demand was likely to prove pretty short-term in nature, it might actually be preferable to simply absorb excess demand in temporarily higher house prices.)

But I don’t see any sign of wide-ranging liberalisation of land-use restrictions in the next few years. As I’ve noted previously, I’m not aware of other countries or major cities that have had tight supply restrictions and materially and sustainably liberalised them. Surely it must come some day, but regulation once established tends to linger for a long time. Import controls, from New Zealand’s history, were another good example.

I don’t think there is any obvious welfare gain for New Zealand in allowing extensive non-resident purchases of houses/apartments if governments also make it hard to bring new urban land to market and utilise it intensively in response to changes in demand. There is none of the technology transfer that might be associated with FDI. There is simply a redistribution – windfall gains to those who happen to own property in Auckland before the demand picked up, and windfall losses to those who would have wanted to purchase in the future. And the gain is simply the result of government-imposed and maintained supply restrictions. In that climate, I see no major problem in principle with some sort of restrictions.

And yet, I remain a little uneasy. In terms of accommodation itself – surely more important than home ownership – it is purchases of houses that are then left vacant that have the stronger adverse effects. Houses that are bought and put back on the rental market maintain the supply of accommodation. And yet we have no data on how important this “left vacant” component might be, and I don’t think the new post-October information requirements will provide any data on this split. Given that demand for house-buying seems relatively price-inelastic, even if the “left vacant” component is itself quite small, as a marginal boost to demand it could still be having quite an impact on price.

Perhaps this is where advocates of the Australian rule (“you can buy, but only a new build”) come in. The Australian rule doesn’t seem to have been very effective, but perhaps a similar one could be much more effectively policed if the authorities were serious about doing so? Perhaps it really is the minimally distortive approach if there is to be new regulation at all? The caveat to that proposition is that if the offshore demand were to prove short-lived we could be left with a nasty over-supply of the sort of housing not overly popular with most New Zealanders. An ample supply of (cheap) apartments sounds good, but real resources will have been diverted into building the properties, skewing the rest of the economy. Real resource misallocation tends to be more costly than changes in asset prices in isolation.  Perhaps that should be less of a concern starting from current house/land prices than in other circumstances?

I’m usually reasonably settled in my views as to appropriate policy responses. For now, on this issue, I’m not. Waiting for October’s data is a convenient line, but I suspect it is a bit of a cop-out. I am left rather closer to Rodney’s 20 per cent stamp duty than I was previously.

I noted in my post yesterday that I was a little surprised at how NBR had characterised differences between Shamubeel Eaqub and me around how to think about the contribution of immigration policy to housing demand. The description was “war of words”, something I didn’t recognise. But I hadn’t seen the article then.

As I noted yesterday, the difference is simply about how to interpret net PLT migration figures. Shamubeel uses them to conclude that immigration policy has not been a major influence on housing demand over 50 years. I pointed out, in response, that immigration policy is about how many non-New Zealanders we let in, and how long we let them stay. It does not affect the movement of New Zealanders at all. Accordingly, if we want to understand the role of immigration policy, we should focus mainly on data on the movement of non-New Zealand citizens. I have used the net inflow of non-New Zealand citizens as a proxy, while noting that it is not a perfect proxy.  On that measure, most of the trend increase in household numbers is now down to immigration policy choices.

So far I thought we just had the sort of difference that crops up all the time in analysing data. Someone proposes a hypothesis, with some numbers, and others respond questioning whether, for example, the data cited are showing quite what the first analyst thought they were. That sort of debate is how we advance our understanding. I didn’t (and don’t) challenge the accuracy of Shamubeel’s numbers (and I’m sure he isn’t challenging mine). The only issue should be what light each set of numbers sheds on the issue (and which issues they each shed light on). As I put it yesterday, if people prefer I’m quite happy to say that (given land use and housing supply restrictions) the large net outflow of New Zealanders has greatly eased pressure on house and urban land prices, and the (even larger) policy-facilitated net inflow of non-New Zealanders has greatly exacerbated pressures.    But only one is a immigration policy matter.

To the extent I had given it any thought, I assumed Shamubeel was approaching the discussion in the same dispassionate way.   In his book (page 129) he notes:

Economist Paul Collier, in his book Exodus argues that we need to talk openly about immigration. Not through the lenses of envy and racism, but in the context of a reasoned and deliberate discussion of why we want immigration, how many people we want, and what kind of people we want.

I nodded, largely agreeing, when I read that passage. Collier’s book is also worth reading.

But this morning I picked up a copy of the print edition of NBR and understood immediately Jenny Ruth’s “war of words” description. Instead of “reasoned and deliberate discussion” my argument is simply dismissed by Shamubeel as “That’s racist”, and “he’s always had this thing about non-New Zealanders. That’s pretty much been the tenor of his work over the last three years”, and “he’s taken a biased approach”.

And there is nothing more than that. There is no sense as to why, as a descriptive exercise, he disagrees with my interpretation of the role of immigration policy in explaining medium-term demand for housing.

Perhaps he provided a more substantive response to the journalist and she didn’t report it, preferring only to report the slurs?  (With apologies to Jenny Ruth) I rather hope so, because Shamubeel is someone whose contributions to economic analysis I have had some time for (indeed on this blog, I encouraged people to read his book) . I think he is much better than is suggested by simply falling back on labels like “racist”, or even “he has this thing about non-New Zealanders”, when someone challenges his framing of the numbers.

I’ve been posing some questions around New Zealand’s immigration policy and the implications for understanding economic performance for five years now. When one is discussing, or researching, the implications of immigration policy, inevitably one is focusing on non-New Zealanders. That is who immigration policy affects. As Shamubeel notes, we (and every country probably) need “reasoned and deliberate discussion of why we want migration, how many people we want, and what kind of people we want”.

In that time I’ve debated the issues and analysis with many people- here and abroad, New Zealand born and foreign born – and I’m pretty sure no one has ever previously accused me of racism.  These are important analytical and policy questions, and the prospects for reasoned and deliberate discussion recede further when people contributing to it the discussion are simply labelled “racist”, rather than engaging on the substance of the issues and analysis.

I hope that Shamubeel will, on reflection, withdraw his slur. As ever, I would be very happy to engage him (or anyone) on the substantive issues (whether interpretative, analytical or policy). And I still think people will benefit from reading his book.

This morning I updated my chart of mortgage approvals per capita. I’ve shown it previously and I like it because it is very timely data.  As you can see, the number of mortgage approvals per capita is running just slightly ahead of last year’s level. But the number of approvals is running below the average, for the time of year, in the 12 year history of the series. In fact, in only three years (2010, 2011, and 2014) have approvals been lower than they are this year.

What do I take from this? First, it reiterates the point that across the country as a whole the housing markets are not particularly buoyant. House sales per capita are far below the peak in the previous boom. Most places have real house prices below previous (2007) peaks, and in most places nominal house prices are pretty flat.

Auckland is a (big) exception. If we had mortgage approvals data on a regional basis, perhaps we’d find that Auckland’s were nearer to or even above the historical median, while the rest of the country might be tracing the lows. But the non-resident demand for houses in Auckland may also be relevant here. We don’t know how large the contribution of that demand to rising house prices is, but much of those purchases (however large or small they are) are probably either cash purchases, or perhaps financed with credit from abroad (not captured in New Zealand data).

I think these data also tend – not conclusively, but suggestively – to confirm the line I ran in my submission last week on proposed investor finance controls. There is simply no evidence that rising house prices in Auckland are to any material extent the result a credit-driven process. If they were, not only should the Reserve Bank be able to point to concrete evidence of deteriorating lending standards, but we would also expect to see lots of mortgages being approved. In fact, only three years have been weaker.

Changing tack, I see that print edition of the NBR has an article about what is described as a “war of words” between Shamubeel Eaqub and me, about the contribution of immigration policy to the house prices pressures in Auckland over time.  I didn’t recognise the description (“war of words”)  and the print edition of NBR isn’t sold this far out in the suburbs so I haven’t yet seen the article, but I just wanted to explain again the difference in how Shamubeel and I are interpreting the same data. I set it out here a few weeks ago.

Shamubeel notes (correctly) that net migration has accounted for only 9 per cent of household formation in New Zealand over the last 50 years. My point is quite simply that net PLT migration (which is probably understated on average over time) is not a description of the contribution of immigration policy. Immigration policy affects the arrival (and duration of stay) of non-citizens. Using the net flow of non-citizens as a proxy for the policy-controlled bit, I concluded that immigration policy accounts for most of the household formation in New Zealand in the last couple of decades. Since 2006, on this proxy, it accounted for 106 per cent of household formation.

I think it would be entirely reasonable to say that, on the one hand, the large net migration outflow of New Zealand citizens over decades has greatly eased pressure on the housing market (taking as given land use restrictions), and, on the other hand, the large net inflow of non-NZ citizens, which is over time a policy-controlled variable, has (even more) greatly exacerbated those pressures. The net effect of the two independent forces might have been modest over 50 years, but one limb is the endogenous behaviour of private NZ citizens, and the other is the direct outcome of policy choices. I might comment further later when I’ve seen the NBR piece.