A week or two back I foreshadowed a forthcoming paper by my former colleague Ian Harrison reviewing the Reserve Bank’s proposals under which the banks would have to greatly increase the volume of capital simply to carry on doing the business they are doing now.

Like me, Ian spent decades at the Reserve Bank.  But much of his time was spent specifically in the area of banking regulation and bank supervision, including leading much of the modelling work done a few years ago as part of the Basle III process, which resulted in something like the current bank capital requirements.   He knows the detail in this area, has consulted on this sort of stuff since leaving the Bank, and has invested a great deal of time and effort over the last couple of months in getting to grips with the Bank’s proposals, reviewing the various papers they’ve published, and going back and reviewing the papers the Bank has cited in support of their case.   The result is his (50 page) review document.   Here are his key conclusions (overlapping in various places with points I’ve made in post here). Ian does not pull his punches.

Part two: Key conclusions

1. The ‘risk tolerance’ approach is a backward step that ignores a consideration of both the costs and benefits of the policy. The soundness test is based on an arbitrarily chosen probability of bank failure that ignores the cost of meeting the target. The Bank has ignored its own cost benefit model which did take the probability of bank failure, the costs of a failure, the interest rate costs of higher capital and societal risk aversion into account.

2. Bank decision based on fabricated evidence. The Banks’s decision to pursue a 1:200 failure target was purportedly based on evidence from a version of the Basel advanced model. It was manipulated to produce the right answer. Initially, a 1:100 target was proposed, but when this couldn’t generate a capital increase, the target was switched to 1:200 at the last minute.

The Bank’s model inputs were not credible. It was assumed that all loans were higher risk business loans and that the probability of loan default, a key model input, was more than two and a half times the estimates the Reserve Bank has approved banks to use in their capital modelling.

The Bank’s analysis was embarrassingly bad, so it attempted to cover this up with a subsequent information paper that was written after the decision was made, and after the Consultation paper was released. It reached the same conclusion on the required level of capital, but only by assuming a 1:333 failure probability, and by using model inputs that were still not credible.

3. A 1:200 target can be met with a capital ratio of around 8 percent. If the Basel model were rerun using credible inputs if would probably show that a 1:200 failure rate can be met with a capital ratio of around 8 percent.

4. The policy will be costly. The Bank has down played the interest rate impact of the policy, saying any increases will be ‘minimal’. Based on its own assessment of the interest rate impact, the annual cost will be about $1.5-2 billion a year. The present value of the cost of the policy could be in excess of $30 billion.

A homeowner with a $400,000 mortgage could be paying an additional $1,000 a year. A business with a $5 million loan could be paying an additional $50,000.

5. The Bank’s assessment that the banking system is currently unsound is at odds with rating agency assessments and borders on the irresponsible. The rating agancies’ assessment of the four major banks is AA-, suggesting a failure rate of 1:1250. The Bank is now saying that, at current capital ratios, the banking system is ‘unsound’ because the failure rate is worse than 1:200. Or in other words the New Zealand banking system is not too far from ‘junk’ status. The international evidence does not support the Bank’s contention that the probability of a crisis is worse than 1:200. The Bank has ignored the fact that banks will need to hold an operating margin over the regulatory minimum, and has not adjusted New Zealand capital ratios to international standards to make a fair like-for-like comparison.

6. The Bank‘s analysis ignores the fact that the banking system is mostly foreign owned. Foreign ownership increases the cost of higher capital because the borrowing cost increases flow to foreign owners. Foreign owners will support their subsidiaries in certain circumstances, which reduces the probability of a bank failure. There is little point in having a higher CET1 ratio than Australia, because if a parent fails then it is highly likely that the subsidiary will also fail, because of the contagion effect. A New Zealand subsidiary might still appear to have plenty capital, but depositors will run and the Reserve Bank and government will have to intervene.

7. The Australian option of increasing tier two capital has been ignored. APRA is proposing to increase bank capital by five percentage points, but will allow banks to use tier two capital to meet the higher target. This provides the same benefits, in a crisis, as CET1 capital, but at about one fifth of the cost. New Zealanders will be required to spend an additional $1.2 billion a year in interest costs for almost no benefit in terms of more resilience to a severe crisis.

8. The benefits of higher capital are modest. Most of the costs of a banking failure are due to borrowing decisions made before the downturn. This will impose costs regardless of the amount of capital held. With current levels of bank capital failures will be rare, with the main cost likely to be a government capital injection. The experience with most banking crises, in countries most like New Zealand, is that governments have recovered most of their costs when the bank shares are subsequently sold.

9. The Bank is mis-selling insurance. The Bank is selling a form of insurance to the New Zealand public, but it vague about the premium costs and has exaggerated the benefits. The premium is the $1.5-2 billion. The benefit would be around a 10 percent reduction in the economic cost of a financial crisis, with an expected return of a few tens of millions.

An informed, rational public would not buy this policy.

10. New Zealand banks already well capitalised compared to international norms. A recent PricewaterhouseCoopers report argued that if New Zealand bank capital ratios were calculated using international measurement standards they would be 6 percentage points higher, placing New Zealand in the upper ranks of well capitalised banking systems. The Reserve Bank critised some details in the report, but has not produced is own assessment as Australia’s APRA has done.

11. The Bank has forgotten about the OBR.   The Open Bank Resolution (OBR) bank failure mechanism, was originally conceived as a substitute for higher capital to reduce fiscal risk, and to reduce the costs of a bank failure. While banks are been required to spend almost $1 billion on outsourcing policies to supportthe OBR, it does not appear in the capital review at all – despite the Governor’s arguments that the main justification for capital increases is to reduce fiscal risk.

The bottom line?

An informed, rational public would not buy this policy.

But, as it happens, an informed rational public won’t get a say. The Governor proposes and (under New Zealand law) disposes: prosecutor, judge, jury, and appellate court in his own case.

Partly, I gather, for his own amusement, and partly to help respond accessibly to some specific assertions/arguments in the more accessible material the Bank has put out to support the Governor’s case, Ian has a separate document, the Pinocchio awards.

The Governor is a great deal smarter and more analytically capable than Donald Trump, but on Ian’s reading, he is resorting to the financial regulator’s equivalent of questionable Trumpian rhetoric to champion the indefensible.  Against Trump there are the courts and Congress.  Against a Governor with a whim and the bit between his teeth……well, nothing really.

It would be interesting to see what the Reserve Bank makes of Ian’s arguments and evidence.

UPDATE: A fairly accessible summary of some of Ian Harrison’s key argument in this article by veteran journalist Jenny Ruth.

 

The Herald’s economics columnist Brian Fallow devoted his column last Friday to the Reserve Bank’s proposal to massively increase the proportion of their balance sheets banks would have to fund by equity.  He ended on what seemed to be a moderately sceptical note.

So we are left trying to weigh a highly debatable but significant cost against an incalculable benefit.

As I’ve noted here, the sort of cost-benefit calculus we can glean from various Reserve Bank publications just doesn’t stack up at all.    In his speech the other day (which a couple of readers have suggested I was too generous towards), the Deputy Governor indicated that the level of GDP might be up to 0.3 per cent lower (permanently) as a result of these changes.    Call it an annual insurance premium of 0.25 per cent of GDP (and bear in mind that the estimates Bascand is using don’t take account of the implications of most of our banks being foreign-owned, which raises the cost to New Zealand).

If we knew with certainty that:

(a) adopting these much higher capital ratios would prevent a crisis in 75 years time that would have otherwise cost 40 per cent of GDP (recall that the proposals aren’t supposed to prevent 1 in 200 year crises, but are presumably supposed to prevent 1 in 150 years crises, and 75 years is halfway through 150 years –  the crisis could happen next year, or in year 149), and

(b) that the new higher capital ratios would be applied consistently for the next 75 years

that might be a borderline reasonable bet –  the benefits might roughly match the costs.

If you thought instead there was no more than a 50 per cent chance that the new policy would be applied consistently for 75 years  (and given how politics, personalities, and conventional policy wisdom has changed quite a bit, to and fro, since 1944 that seems generous) , you’d need to be preventing a crises twice as large for the insurance policy to be worthwhile.

As a benchmark for how serious these events can be, you could think about the United States since 2008.   Over 2008/09, the United States experienced the most severe banking crisis of (at least) the modern era, since (say) the creation of the Federal Reserve.   (The Great Depression itself was, of course, worse, but it was mostly a failure of monetary policy management and institutional design rather than a banking crisis).

I’ve shown this chart previously.  Real GDP per capita for the US and New Zealand from 2007q4.

The United States was the epicentre of the systemic banking and financial crisis.  New Zealand did not experience a systemic banking crisis at all.  At least on the OECD’s estimates, the output gaps of the two countries were pretty similar in 2007.   And there is no just credible way you can come up with estimates that have the United States doing (cumulatively, in present value terms) 40 per cent of GDP worse than New Zealand.   Do the comparisons of the US with other OECD countries that didn’t run into serious banking problems –  eg Australia, Canada, Norway, Israel, Japan –  and you still won’t find cumulative output losses sufficient to justify the sort of tax the Reserve Bank of New Zealand wants to lump on our economy.

In fact, for what it is worth, here is the cumulative labour productivity (real GDP per hour worked) growth of the United States and those non-crisis countries for the period 2007 to 2017 (from the OECD databases).

United States	10.6
Australia	14.0
Canada	9.6
Israel	10.1
Japan	8.3
Norway	3.0
New Zealand	4.4

Sure, there are all sorts of other things going on in each of these countries.   But bad as the 2008/09 banking crisis was in the United States, it is all but impossible to come up cumulative cost estimates that would reach even 20 per cent of GDP (crisis country experience over the experience of non-crisis countries).

And if the Reserve Bank continues to believe otherwise, the onus should be on them to make the case, to set out their arguments, assumptions, and evidence.

(Note, that none of this is to suggest that the 2008/09 recessions were anything other than undesirable and costly.  Recessions generally are.  But there are lots of low probability bad things in our world, and not all of them are worth paying the price to try to prevent.)

But even having come this far, I’ve risked conceding too much to the Reserve Bank ‘s story.  In part, that is because the storytelling implicitly treats higher capital ratios as sufficient to spare an economy all the costs associated with a banking and financial crisis.  And in part (but not unrelated to the previous point) because it tends to treat systemic banking crises as bolts from the blue –  unlucky bad draws imposed on a country when the Lotto balls are selected.   Neither is true.

Unfortunately, Brian Fallow seemed to buy into some of this sort of implicit reasoning in his column.

But in a complicated and perilous world it is idle to suggest that stress tests or banks’ internal modelling can accurately quantify the probability and magnitude of a major economic shock that could threaten the solvency of banks. Or even forecast the nature of the shock: another global financial crisis, perhaps?

The last one inflicted the deepest recession New Zealand had suffered since the 1970s.

A pandemic against which we are pharmacologically defenceless? It’s 100 years since the Spanish flu killed more people than WWI.

International hostilities breaking out in some surreptitiously weaponised but systemically vital realm of cyberspace?

Or a more conventional geopolitical conflict? The current crop of world leaders do not inspire confidence.

One might note first that the stress tests the Reserve Bank requires banks to undertake have deliberately used highly adverse combinations of shocks (rising unemployment, falls in house prices etc).  They are deliberately designed to look at really adverse events.

One might also note that –  contrary to the implication here and in Geoff Bascand’s speech –  the banking crisis (failure of DFC, need to recapitalise BNZ) did not “inflict the deepest recession New Zealand has suffered since the 1970s”, rather it was one (probably rather modest) factor in a range of contributors (disinflation, structural reform, fiscal adjustment, downturns in other countries etc).

But what I really wanted to focus on were the final three paragraphs in that extract, which imply that banking crises arise out of the blue.  And they just don’t.  They never (or almost never) have.  Might they in future?  Well, I suppose anything is possible, but we can’t sensibly take precautions against everything.

The prospect of a serious pandemic is pretty frightening.  But awful as the episode 100 years ago was, it didn’t result in systemic banking crises.    Cyber-warfare sounds pretty scary as well, but it isn’t remotely clear how higher bank capital requirements protect us against that.   World War Two was dreadful, but it didn’t result in systemic banking crises either.    There was a serious financial crisis at the start of World War One, but it was a liquidity crisis not a solvency one, and capital requirements (then, or in some future unexpected outbreak of hostilities) are pretty irrelevant in a crisis of that sort.    Do the leaders of the world as individuals or a group command much confidence?  Well, no, but then in the last 100 years or so periods when it was otherwise seem rare enough  (the PRC and USSR border war of 1969 anyone)?  I suppose if we go back far enough a repeat of the Black Death –  wiping out a third of the population –  wouldn’t be great for the value of bank collateral, but (a) capital proposals aren’t supposed to counter 1 in a 1000 year shocks, and (b) you might think there would be bigger things to worry about then (and a high likelihood of statutory interventions to redistribute gains and losses anyway).

Systemic banking crises –  one where banks and borrowers lose lots and lots of money and 5 per cent, 10 per cent or more of bank loans are simply written off in a short space of time –  simply do not arise out of the blue, as decent well-managed banks and universally responsible borrowers are suddenly hit by some totally unforeseeable event.  Rather they are –  always and everywhere – the outcomes of choices made over the years (typically only a handful) previously.    Lenders and borrowers make bad –  wildly overoptimistic  –  choices, some just going along for ride, others actively pushing the envelope.  In the process, real resources in the economy in question are misallocated, perhaps quite badly, but that cost is something that is really only apparent (only crystallises) when the boom comes to end, whether it ends in a whimper or a bang.   Sometimes government policies can play a direct part in helping to generate the mess.  In the United States, for example, the heavy state involvement in the housing finance sector greatly exacerbated the imbalances that crystallised in 2008/09.  In Ireland, for example, entering the euro and taking on the interest rate fit for Germany and France when Ireland might have been better with New Zealand interest rates, was a big part of the story.  Fixed exchange rates have often been a significant factor, both giving rise to initial imbalances, and aggravating the difficult resolution afterwards).  Transitions out of periods of heavy regulation can play a role too: in New Zealand (and various other countries) in the late 80s, neither lenders nor borrowers really had much idea of operating in a liberalised financial system.  You can go through every modern financial crisis –  and probably plenty of the older ones too – and point to the excesses, the over-optimistic lending and borrowing that accumulated in the preceding years.  Japan –  crisis of the 1990s – was yet another prominent example (Zambia in 1995, a crisis I was quite involved in, was the same).

Would higher required capital ratios have prevented any (many) of these episodes?  One can’t answer with certainty.  They might have prevented some bank failures themselves, but that isn’t the issue I’m focused on here (which is about the bad lending/borrowing in the first place).    Perhaps a few banking systems really went crazy because creditors thought they could offlay all the risks on the state, but that isn’t a compelling story more generally.  After all, shareholders still stood to lose everything.   A more plausible interpretation (to my mind) of those periods of undisciplined lending/borrowing is that people simply misjudged the opportunities, and got carried away by excess optimism, in ways that meant they just pay much less attention to the potential downsides.  The world was different (so people were being told, or they told themselves).  If so, most of the misallocations of real resources would happen quite independently of the levels of bank capital requirements that were imposed.  And you can mount an argument that high capital requirements may tend to encourage banks to seek out more risks than they would otherwise take, especially in buoyant times, concerned to keep up rates of return on equity.    Even if that doesn’t happen, disintermediation would see more risks to be taken on in the shadows –  no less resource misallocation in the process, but rather less visibility to the authorities and resolution agencies.

Perhaps good and active bank supervision can prevent those excesses, and misallocations, building up.  But that is a (very) different case from one in which ever-more-demanding capital buffers  supposedly eliminate (or reduce to minimal levels) the costs when the bad lending crystallises.  It isn’t a case the Reserve Bank has made –  and they probably wouldn’t, as historically they have been (rightly) fairly sceptical about the value hands-on supervision can add.  And it is a case I am pretty sceptical of.   And for which there is not much evidence (even allowing for the fact that crises avoided tend to be not very visible).   Much as APRA likes to suggest it is an example (in the 2000s) I don’t think they were really tested.

Bank supervisors –  and their bosses in particular –  breath the same air as everyone else in a society, and when lending and borrowing in a particular country is going badly off course, it is unlikely that the bank supervisory agency will be taking (for long) a very different stance from those around them, and those who appoint them.   This isn’t an argument about corruption –  although regulators perceived to be realistic and responsive will no doubt attract a better class of well-remunerated job offer –  but about political and economic realism.  But even if better Irish regulators (say) could really have made a difference in the 2000s, what was really needed were different lending/borrowing practices, not just more capital.  More capital wouldn’t have avoided a nasty aftermath (even if it would have reduced some fiscal costs) –  and at the peaks of self-confident booms, capital is cheap and easy to raise.

And so we are brought back to the specifics of New Zealand where:

• repeated stress tests conclude that our banking system is resilient to very very nasty shocks,
• the Reserve Bank tells us at every FSR that the financial system is strong and sound,
• we have control of our own monetary policy, including a floating exchange rate (34 years today),
• we have healthy public finances,
• we have little active involvement of the government is directing finance

Against that backdrop, and against the experience in which it is hard to conclude that banking crises themselves (as distinct from the bad lending that later gave rise to them) are worth more than a few percentage points of GDP, in a country not prone to systemic financial crises (the episode doesn’t even meet the test in many collections of crises), with our banks owned mostly from a country also with no track record of frequent serious financial crises, the case just hasn’t been at all convincingly for compelling banks to fund so much more of their balance sheets with equity.  Doing so will, on the Reserve Bank’s own telling, involve real economic costs.  For benefits that are, at best, tiny and far-distant.

It is disconcerting that in none of the Reserve Bank’s material do they ever show signs of engaging with any of this sort of analysis.  Instead, they have a policy preference and prefer assertions and flimsy analysis to any serious engagement with the issues and experience.

My former colleague Geof Mortlock has another piece on interest.co.nz making the case for splitting up the Reserve Bank and creating a separate Prudential Regulatory Agency.  I strongly agree with him on that –  and have argued the case here last year.  Geof is probably more optimistic than I would be about what bank regulators can add, but on this particular item we seem to be as one.   Among his long list of what is wrong with the Bank’s conduct of its financial regulatory functions, introduced thus

Geof Mortlock argues the Reserve Bank is about as much use as a financial regulator as is a cricket umpire who is nearly blind and who understands little about the game

is this:

the recent release of bank capital regulation proposals that would see banks in New Zealand being required to hold a very high level of capital compared to other countries, with potentially adverse consequences for borrowers’ access to credit, an increase in interest rates and adverse impacts on the economy – and all on the basis of shockingly flimsy analysis by the Reserve Bank.

Quite.

 

 

The Reserve Bank Deputy Governor’s speech yesterday was released under the title Safer banks for greater wellbeing, while the handout at the venue went even further and was headed (in a very big font indeed) Safer banks = safer society.    Count me sceptical.

It was a disaappointing speech.  Plenty of people turned up to the university at lunchtime, including such eminent figures as the Governor and the former Deputy Governor (Grant Spencer), but we were treated to something not much more than the ECON101 case for huge increases in bank capital requirements.  Geoff Bascand’s speeches have typically been the most thoughtful and considered of those given by Reserve Bank senior management.  This latest effort didn’t reach that standard.  Instead we had alarmist rhetoric about history, key charts deployed for support rather than illumination, and no attempt to dig deeper and use whatever that digging might throw up to shed light on the case the Bank is making (in a cause in which it is prosecutor, judge, and jury in its own case).

History first.  As Bascand noted, New Zealand hasn’t had much history with systemic financial crises (although there is an interesting article here on the two episodes we have had).  The first was in the 1890s, culminating in the bailout (and partial nationalisation) of the BNZ in 1894 (and the fiscal cost of that bailout (per cent of GDP) was a bit larger than in the more recent BNZ bailouts).  Bascand really only notes this episode in passing but here is the chart of (estimated) GDP per capita during that period.

It was certainly a nasty recession –  in an era when economies were more volatile than they are now –  but it didn’t last long, and even if you attributed all the lost output to the financial crisis itself (and none to the misallocation of resources and bad lending that led to the banking problems) you only end up with total lost output of around 10 per cent of GDP.   And that in a regime in which the exchange rate was fixed and New Zealand had no discretionary control of interest rates.   (The 1890s crisis in Australia would have provided much stronger superficial support for Bascand’s argument, but with the same attribution issues.)

The more recent episode was involved two recapitalisations of the BNZ (and the failure of DFC, the travails of NZI Bank etc) in the late 1980s and early 1990s.  Bascand notes that he lived through this period as a Treasury official and goes on to say

If you ask someone who’s lived through a banking crisis, they’ll likely tell you that the impacts were not only significant, but lasting. Perhaps the person you talk to may have lost their job as a result of the crisis, and if not, it might have been their spouse, a friend, or a neighbour. Maybe you speak to a young couple that had purchased their first home just prior to the crisis, only to see its value decline by 30% in the months following the crisis, forever altering their outlook on the economy and their willingness to make another significant investment. Or maybe you speak to someone who just graduated from university prior to the crisis, only to enter a depressed labour market, and forced to accept work well below their educational qualifications and abilities, forever altering their desired career path.

Talk to these people, and I think they will tell you that banking crises have altered their lives in ways they wished it hadn’t. I think they will also tell you that banking crises should not be accepted as an unavoidable fact of life.

For those that lived through the recession we experienced here in the early 1990s, you will recall that some industries were decimated, and a generation of workers lost. Many of these workers were not able to re-enter the workforce easily and lost valuable skills while trying to find suitable employment. And while recessions sometimes occur in the absence of a banking crisis, it is common for banking crises to ultimately result in recessions.

Actually, most recessions (not just “sometimes”) don’t involve banking crises, and it is asserting that which needs to be proved to suggest that banking crises “result in” recessions.  Yes, banking crises often happen at the same time as recessions.    Initial waves of bad lending, over-optimism, and misallocated lending often contribute to both the economic downturn and to the banking sector problems.   Big increases in capital ratios from already high levels won’t change any of that.  Quite possibly any disruptions to the intermediation process associated with banking failures (or near failures) exacerbate the economic downturn, or slow the subsequent recovery,  but the Bank cites no studies (and I’ve not seen any) that attempt to separate out those effects.  Implicit in a lot of this is handwaving around the poor global economic performance in the last decade, when countries that haven’t had financial crises have (on average) not performed much better than those that have.

And, of course, in the New Zealand in the early 1990s there was a great deal else going on.   Although he doesn’t do so in the text, in his address Bascand did acknowledge that point, but simply acknowledging the point in passing –  while talking at the same time of 11 per cent unemployment – isn’t really enough.    We had the combined effect of:

• disinflation (getting inflation down from 10-15 per cent to something in the 0-2 per cent range),
• significant fiscal adjustment (recall the large deficits at the end of the Muldoon term),
• far-reaching structural reforms in the New Zealand public sector, including the new SOEs, that involved laying off lots of workers,
• significant reductions in trade protection,
• and the after-effects of an asset price and commercial property boom, with considerable misallocated resources (all of which had occurred fresh out of liberalisation, when neither borrowers nor lenders –  let alone regulators – really knew what they were doing, what the relevant parameters and possibilities of the new market economy might have been.  In the aftermath, whatever happened to the supply of credit, there wasn’t much demand for it either.

So I’m quite happy to believe that the banking crisis itself may have had some economic costs, but if the Bank wants to argue that they were more than a small fraction of overall costs of that period the onus is surely on them to produce the research in support.  As it is, (and despite paying little attention at the time to potential financial intermediation channels) the Reserve Bank’s forecasters were surprised by the speed of the economic recovery from the 1991 recession.  But I guess it is easier to simply fling round emotion-laden rhetoric about mental health etc.

And even narrowing things down to the BNZ problems, it is worth keeping that episode in perspective.  The paper I linked to earlier records that the recapitalisation of the BNZ cost around 1 per cent of GDP.  Better never to have had to do it, but that is pretty small by the standards of serious systemic banking crises (and, as I understand it, the direct outlay was fully recouped later).  Perhaps relevantly to this debate, I was tempted to ask Bascand yesterday if he had any idea what risk-weighted capital ratio the BNZ would have had in the late 80s.    Hard to estimate without someone doing some very detailed research, but I talked to someone else who was around at the time who estimates that at present (before the latest Reserve Bank proposals) the BNZ would be at least twice, possibly three times, better capitalised now than it was then.  But of course you get none of this flavour from Bascand’s speech, or from any of the Reserve Bank documents published in recent months.

The Bank’s stress tests didn’t get a mention in the speech but a questioner asked about them.    Bascand attempts to parry the question noting that they were “slightly artificial constructs” (sure, and so are any analytical techniques) but offered, without further prompting, that they certainly suggested “pretty resilient banks”.  Nothing was offered in elaboration as to why, if severe stress tests show that banks not only don’t fail they don’t even fall below existing minimum capital ratios, regulators should be so insistent on such large further increases in the required capital ratios.  I guess it is a bit awkward for them, and silence is easier than explanation?   (Incidentally, the same questioner asked if much higher capital ratios would have some quid pro quo in lower supervisory intensity, but Bascand declared that not only would capital ratios be increased but that the Bank will increase its supervisory intensity.)

One of the areas the Bank has been pushed on is how their proposals compare to what is being done in other advanced countries,  They’ve still given no satisfactory answers, not even something as (apparently) simple as an indication of the all-up expected capital ratios (core equity and total) APRA will expect for the Australian banking groups.  An apparently knowledgeable commentator here has suggested that the total capital requirements are likely to be similar, but that the Reserve Bank is insisting on a much larger share of that being made up of (expensive) common equity.    If true, that would be useful context for evaluating the Bank’s proposals.  It is the sort of information they should have presented when the proposal was first released, more than two months ago now.

In the speech itself, Bascand included a couple of charts/tables intended to support his view.   The first was this one (I’ve added the circling), included in the speech with no elaborating comment at all.

The table is taken from a 180 page paper, and is supposed to represent an estimate of where banks in other countries will get to when the Basle III standards are fully phased in.  It isn’t clear –  from the speech or from skimming through the 180 pages, although I presume there is a simple answer –  whether these numbers are minimum required capital ratios or forecast actual capital ratios.

I’ve highlighted the numbers for the 75th percentile for the Group 1 banks (which includes the Australian parent banks) and the globally systemically significant (GSIB) subset of those.    The Reserve Bank’s current proposals will require the four largest New Zealand banks to have minimum capital ratios of 16 per cent of risk-weighted assets.  Actual capital ratios –  and it is actual capital ratios that provide the buffer not minima –  will be higher again.   These are higher than the 75 percentile for the world’s biggest and most problematic (if anything goes wrong) banks.  The G-SIB banks are typically complex, and cross multiple national boundaries, and there is no clear or robust idea how any potential failure will be resolved.   On any sensible framework you would suppose that minimum capital requirements for such banks would be materially higher than those for vanilla retail banks operating in a single country, with large and strong parents.  But not, it seems, to the Reserve Bank of New Zealand.

And, as it happens, this table doesn’t help us with one of the biggest differences between the way New Zealand capital ratios have been calculated and those in many European countries (in particular).    The minimum risk weights here are generally accepted to be materially higher than those applied in many other advanced countries.  Using the same sorts of risk weights used in many other countries, the capital ratios of our banks would appear quite a bit higher.

How much higher?   Well, a couple of papers the Reserve Bank itself released (here and here) commenting on some PWC analysis shed light on that.    Take the Australian situation first.  PWC did some work there which concluded that Australian risk-weighted capital ratios were understated by 4 percentage points.  APRA didn’t agree.  They did their own study and concluded that the difference was “more like 3 percentage points”.  That is stilll a big difference.  PWC’s work on New Zealand concluded that the difference here was more like 6 percentage points.  The Reserve Bank  didn’t do its own study, but the internal note they did do concluded

….even after correcting for these biases, there may well continue to be a degree of reported conservatism, such that while we do not have much confidence in the 600 basis point figure they reach, we would accept the overall assessment that we are likely to be more conservative than many of our peers;

Since minimum risk-weights imposed by the Reserve Bank were typically higher than those imposed by APRA, it would seem unlikely that the difference here is less than the 3 percentage points APRA accepted in their study.

And much of this carries over to the new Reserve Bank capital proposals.  Among its plans, the Bank is proposing to use a floor such that the big banks (using their internal models) cannot have capital ratios less than 90 per cent of what would be generated if the standardised approach (applying to other banks) were applied to their portfolios.  That is one of the changes that looks broadly sensible to me.  But apparently most other advanced countries are planning to use a floor of 72 per cent.   All else equal, a 16 per cent capital ratio calculated on Reserve Bank rules could easily be equivalent to something like 19 per cent in many other countries’ systems.   And not even the 95th percentile of G-SIB banks will –  according to the BCBS table –  have a Tier 1 capital ratio of 19 per cent.

I quite accept the Deputy Governor’s point that doing international comparisons well is hard.    But the Reserve Bank has a lot more resources, including membership of international networks of regulatory agencies, than most people reacting to their proposals.  And yet they’ve made little or no effort to engage in robust, open, benchmarking against what other countries are doing –  not even Australia, when resolution of any problems in the big 4 banks will inevitably be a trans-Tasman affair.

The Deputy Governor then included another chart, with not much more comment

It certainly looks helpful to the Reserve Bank’s case, suggesting that current capital ratios (calculated this way) for big New Zealand banks are currently low by international standards and would still be not-high if the new proposals were applied (the Bank assumes quite a small margin of actual capital over minimum required – for reasons that have some plausibility).

But one needs to dig behind this chart and see what is going on.  The rating agency S&P engages in its own attempt to calculate risk-weighted capital ratios for a large number of banks, using its own risk-weighting framework.   But a great deal depends on the “economic country risk score” the S&P analysts assign.    And they take a dim view of New Zealand, assigning us a score of 4 (on a 10 point scale).  Here is what that means for housing risk weights

And there are similarly large differences for the corporate risk weights.

As I said, S&P gives New Zealand a 4.   But Sweden, Norway, Belgium, Switzerland, and Canada all get a 2.    You might think there are such large systematic economic risk differences between New Zealand and those countries, but I doubt the Bank really does, and I certainly doubt. I wrote about this a few years ago where I noted

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.

Score us as a 3 or even a 2 and suddenly the Deputy Governor’s chart will have the implied capital ratios for New Zealand banks a lot higher.

There aren’t easy right or wrong answers to some of these issues, but the uncertainties just highlight how much better it would have been if the Reserve Bank had engaged in an open consultative process at a working technical level, before pinning their colours to the mast with ambitious far-reaching proposals.      As another marker of what is wrong with the process, the Deputy Governor told us yesterday that the Bank will be releasing an Analytical Note on the Bank’s estimates of the costs of their proposals: it will, we were told, be out in a “couple of weeks”, by when two-thirds of the (extended) consultative period will have passed.

In the question time yesterday, the Deputy Governor was given the opportunity by a sympathetic questioner to articulate why the Bank should be conservative relative to many other overseas banking regulators.   He didn’t offer much: there was a suggestion that New Zealand is particularly subject to shocks, and a claim that New Zealanders are strongly risk-averse (but not evidence, let alone that these preferences are stronger than those of people in other advanced countries).  I can identify grounds on which some regulators might sensibly be more conservative than the median:

• if you were in a country with a bad track record of repeated financial crises.  But that isn’t New Zealand,
• if you were in a country where much of credit was government-directed (directly or through government-owned banks).  But that isn’t New Zealand.
• if you were in a country that depended heavily on foreign trade and yet had a fixed nominal exchange rate. But that isn’t New Zealand.
• or no monetary policy capability of its own. But that isn’t New Zealand.
• or if you were in a country where the public finances were sick.  But that isn’t New Zealand,
• or if you were in a country where the big banks were very complex and you weren’t confident you understood the instruments. But that isn’t New Zealand.
• or if you were in a country where the big banks had no cornerstone shareholder, were mutuals, or where the cornerstone shareholder was from a shonky regime. But that isn’t New Zealand.

The case just doesn’t stack up.

And, as I noted yesterday, using the numbers the Deputy Governor himself cited, a simple cost-benefit assessment doesn’t seem to stack up either.  We are asked to give up quite a lot of income (PV of $15 billion on his numbers) for some wispy highly uncertain probability of easing a recession in perhaps 75 years time.

If there is a robust case for what they want to do, it just hasn’t yet been made.

At lunchtime I went to hear Reserve Bank Deputy Governor Geoff Bascand make the case for his boss’s proposal to require the locally incorporated banks operating in New Zealand to fund a much larger proportion of their balance sheets with equity capital.  I will write tomorrow about a range of other points that were, and weren’t, made.  But for now I wanted to pick up just one number he used in making the case.

In the course of his presentation, Bascand used a slide which reported the Bank’s view that these changes in capital requirements will lower the long-run level of GDP by a bit less than 0.3 per cent.  I hadn’t seen the number before (maybe it was in the documents, in which case I missed it), but what struck me was Bascand’s suggestion that this is “not a very big number”.

Looked at quickly, perhaps that is true.  But it is a price the economy will have to pay each and every year.  Using the standard Treasury discount rate (6 per cent real), the present value of those costs is about 5 per cent of one year’s GDP ($15bn or so in today’s money).   The precise number isn’t certain –  could be less, could be more – but whatever the cost, we are stuck with it, year in year out, for as long as this policy proposal was in place.

And what are getting in return for our lost $15 billion?

And that is where things get very uncertain.  The Bank will tell us that we are avoiding the terrible costs of a financial crisis.  They will quote various numbers at you, but on this occasion Geoff Bascand included a slide in which a typical advanced country financial crisis had a cumulative economic cost (lost output) of 23 per cent of GDP.

But even if one uses that number as a starting point, an increase in capital ratios of the sort the Bank proposes aren’t going to save all that lost output because:

• as I’ve noted repeatedly, much of any output loss associated (in time) with a financial crisis is the result of the bad lending and misallocation of real resources that may have led to the crisis, but did not result from it. It would happen anyway. We don’t know what the right split is –  as I noted yesterday, I’m not aware of any papers that really make the attempt –  and the Reserve Bank hasn’t told us its estimate, and
• we aren’t starting from near-zero capital, but from actual capital ratios that even the Bank concedes are relatively high by international standards at present, and
• even these capital requirements are not supposed to spare us from all crises, just keep them to no more than 1 in 200 years.

It is the additional reduction in output losses (not the total loss) resulting from these  capital proposals that has to be compared to the annual output loss (the “insurance premium” if you like) of simply putting the policy in place.

As I noted yesterday, the policy proposals aren’t supposed to protect us from a 1 in 200 year crises, but they should protect us from, say, a 1 in 150 year crisis.   Perhaps we –  generously in my view, on my reading of the historical experience –  take the view that the further increase in capital requirements can save us from a 10 per cent of GDP loss when the crisis happens.

We don’t know when in the 150 years the actual crisis will happen, so lets assume that it happens in year 75 (half way through).    We could discount back that saving –  10 per cent of GDP 75 years hence – at a 6 per cent discount rate and the resulting present value is about 0.15 of GDP.   In other words, the present value of what we save –  that quite severe event, but a very long way in the future – is a bit less than one year’s insurance premium.

Another way of looking at that number is to take the 10 per cent of GDP (not) lost and spread it out over 150 years.   That becomes an annual saving of 0.06 per cent of GDP.  In exchange for which we pay a premium of getting on for 0.3 per cent of GDP.   It would take a future crisis event hugely more costly to make the insurance even remotely worthwhile.

And all that assumes we know that we’ll actually protect ourselves.  But we don’t.  Up front, we know that the banks at present are pretty strong (as even the Reserve Bank acknowledges, and that is what the stress tests show).  There is no chance that this really severe crisis will happen in the next few years.   And, on the other hand, there is no pre-commitment mechanism to guarantee that the new capital requirements are kept in place for 50, 75, or 150 years.  No pre-commitment mechanism, and no probability either –  just look at how often regulatory rules change, in this and many other areas.

And while the Reserve Bank’s GDP loss numbers are about long-term levels, there is also the transition to consider.  Most probably, in the course of the transition credit will be less readily available.  Most probably, during the transition the next recession will occur (not because of the policy change, but just the passage of time and accumulation of external risks),  and in that environment banks seeking to pull back on credit or widen margins are likely to result in a bit more of output cost than the long-term estimate.

In other words, if the Bank goes ahead with this proposal, we will be poorer by up to 0.3 per cent per annum for each and every year the new rules are in place.  There will, most likely be some additional losses in the transition period.     And to gain what?   Basically nothing in the next few years – lending standards have been sufficiently robust there is no credible way over that period banks will run through existing capital over that horizon, let alone the new higher levels.  And beyond that, the annualised gain (or PV of a lump sum saving decades ahead) is just tiny on plausible estimates of the marginal GDP savings higher capital ratios might one day deliver us.

To sum up, there are certain to be annual costs, exacerbated in a transition.  There is no certainty future Governors will stick to the policy even if it is adopted this year (if they don’t we will have paid the premium and got nothing), and even if they do it would require incredible (ie literally unbelievable) future GDP savings – in the event of a far-distant crisis –  to make paying the insurance worthwhile.

0.3 per cent per annum –  in a country struggling for all the productivity it can get – might look like “not a very big number”.   But the protection it purports to buy us looks to be of derisory, and highly uncertain, value.  Against that backdrop, the (capitalised) $15 billion price tag could be spent on a lot more worthy things.   The Deputy Governor’s speech attempts to tie the Bank onto the wellbeing bandwagon (“Safer banks for greater wellbeing”).  Well, you can buy a really large amount of, say, mental health services (to take a theme from this morning’s Herald – and from Bascand’s speech) with a $15 billion lump sum.