In the press release for last week’s Reserve Bank Financial Stability Report, the Governor commented that

Our preliminary view is that higher capital requirements are necessary, so that the banking system can be sufficiently resilient whilst remaining efficient. We will release a final consultation paper on bank capital requirements in December.

In commenting briefly on that, I observed

Time will tell how persuasive their case is, but given the robustness of the banking system in the face of previous demanding stress tests, the marginal benefits (in terms of crisis probability reduction) for an additional dollar of required capital must now be pretty small.

There wasn’t much more in the body of the document (and, as I gather it, there wasn’t anything much at the FEC hearing later in the day), so I was happy to wait and see the consultative document.

But the Governor apparently wasn’t.  At 12.25pm on Friday a “speech advisory” turned up telling

The Reserve Bank will release an excerpt from an address by Governor Adrian Orr on the importance of bank capital for New Zealand society.

It was to be released at 2 pm.   The decision to release this material must have been a rushed and last minute one –  not only was the formal advisory last minute, but there had been no suggestion of such a speech when the FSR was released a couple of days previously.

And, perhaps most importantly, what they did release was a pretty shoddy effort.    We still haven’t had a proper speech text from the Governor on either of his main areas of responsibility (monetary policy or financial regulation/supervision) but we do now have 700 words of unsubstantiated (without analysis or evidence) jottings on a very important forthcoming policy issue. which could have really big financial implications for some of the largest businesses in New Zealand and possibly for the economy as a whole.

The broad framework probably isn’t too objectionable.  All else equal, higher capital requirements on banks will reduce the probability of bank failures, and so it probably makes sense to think about the appropriate capital requirements relative to some norm about how (in)frequently one might be willing to see the banking system run into problems in which creditors (as distinct from shareholders) lose money.  At the extreme, require banks to lend only from equity and no deposit or bondholder will ever lose money (there won’t be any).

But what is also relevant is the tendency of politicians to bail out banks.  Not only does the possibility of them doing so create incentives for bank shareholders to run more risks than otherwise (since creditors won’t penalise higher risk to the same extent as otherwise), but there is a potential for –  at times quite large –  fiscal transfers when the failures happen.  Politicians have more of an incentive to impose high capital requirements on banks when they acknowledge their own tendencies to bail out those banks.  If, by contrast, they could resist those temptations –  or even manage them, in say a model with retail deposit insurance, but wholesale creditors left to their own devices –  it would also be more realistic to leave the question of capital structure to the market –  in just the same way that the capital structure of most other types of companies is a mattter for the market (shareholders interacting with lenders, customers, ratings agencies and so on).

But nothing like this appears in the Governor’s jottings.  Instead, we have the evil banks, the put-upon public and the courageous Reserve Bank fighting our corner.   I’d like to think the Governor’s analysis is more sophisticated than that, and one can’t say everything in 700 words, but…..it was his choice, entirely his, to give us 700 words of jottings and no supporting analysis, no testing and challenging of his assumptions etc.

There are all manner of weak claims.  For example

We know one thing for sure, the public’s risk tolerance will be less than bank owners’ risk tolerance. 

I think the point he is trying to make is about systemic banking crises –  when large chunks of the entire banking system run into trouble.  There is an arguable case –  but only arguable –  for his claim in that situation, but (a) it isn’t the case he makes, and (b) I really hope that (say) the shareholders of TSB or Heartland Bank have a lower risk tolerance around their business than I do, because I just don’t care much at all if they fail (or succeed).  Their failures  –  should such events occur –  should be, almost entirely, a matter for their shareholders and their creditors, with little or no wider public perspective.

There are other odd arguments

Banks also hold more capital than their regulatory minimums, to achieve a credit rating to do business. The ratings agencies are fallible however, given they operate with as much ‘art’ as ‘science’.

Bank failures also happen more often and can be more devastating than bank owners – and credit ratings agencies – tend to remember.

And central banks and regulators don’t operate “with as much ‘art’ as ‘science'”?   Yeah right.   And the second argument conflates too quite separate points.  Some bank failures may be “devastating” –  although not all by any means (remember Barings) –  but the impact of a bank failure isn’t an issue for ratings agencies, the probability of failure is.   And I do hope that when he gets beyond jottings the Governor will address the experience of countries like New Zealand, Australia, and Canada where –  over more than a century –  the experience of (major) bank failure is almost non-existent.

The Governor tries to explain why public and private interests can diverge (emphasis added)

First, there is cost associated with holding capital, being what the capital could earn if it was invested elsewhere. Second, bank owners can earn a greater return on their investment by using less of their own money and borrowing more – leverage. And, the most a bank owner can lose is their capital. The wider public loses a lot more (see Figure 2).

But what is Figure 2?

figure 2

Which probably looks –  as it is intended to –  a little scary, but actually (a) I was impressed by how small many of these numbers are (bearing in mind that financial crises don’t come round every year), and (b) more importantly, as the Governor surely knows, fiscal costs are not social costs.  Fiscal costs are just transfers –  mostly from one lots of citizens (public as a whole) to others.  I’m not defending bank bailouts, but they don’t make a country poorer, all they do is have the losses (which have arisen anyway) redistributed around the citizenry.  If the Governor is going to make a serious case, he needs to tackle –  seriously and analytically –  the alleged social costs of bank failures and systemic financial crises.  So far there is no sign he has done so.  But we await the consultative document.

There is a suggestion something more substantive is coming

We have been reassessing the capital level in the banking sector that minimises the cost to society of a bank failure, while ensuring the banking system remains profitable.

The stylised diagram in Figure 3 highlights where we have got to. Our assessment is that we can improve the soundness of the New Zealand banking system with additional capital with no trade-off to efficiency.

and this is Figure 3

capital chart

It is a stylised chart to be sure, but people choose their stylisation to make their rhetorical point, and in this one the Governor is trying to suggest that we can be big gains (much greater financial stability, and higher levels (discounted present values presumably) of output) by increasing capital requirements on banks.

I don’t doubt that the Bank can construct and calibrate a model that produces such results.  One can construct and calibrate models to produce almost any result the commissioning official wants.  The test will be one of how robust and plausible the particular specifications are.  We don’t know, because the Governor is sounding off but not (yet) showing us the analysis.  Frankly, I find the implied claim quite implausible.   Probably higher capital requirements could reduce the incidence of financial crises.  But the frequency of such events is already extremely low in well-governed countries where the state minimises its interventions in the financial system, so I don’t see the gains on that front as likely to be large.    And, as I’ve outlined here in various previous posts I don’t think that the evidence is that persuasive that financial crises themselves are as costly as the regulators (champing at the bit for more power) claim.  And many of the costs there are, arise from bad borrowing and lending, misallocation of investment resources, which are likely to happen from time to time no matter how well-capitalised the banks are.

There are nuanced arguments here, about which reasonable people can disagree. But not in the Governor’s world apparently.

He comes to his concluding paragraph, the first half of which is this

A word of caution. Output or GDP are glib proxies for economic wellbeing – the end goal of our economic policy purpose. When confronted with widespread unemployment, falling wages, collapsing house prices, and many other manifestations of a banking crisis, wellbeing is threatened. Much recent literature suggests a loss of confidence is one cause of societal ills such as poor mental and physical health, and a loss of social cohesion.

Oh, come on.  “Glib proxies”……..    No one has ever claimed that GDP is the be-all and end-all of everything, but it is a serious effort at measurement, which enables comparisons across time and across countries.  Which is in stark contrast to the unmeasureable, unmanageable, will-o-wisp that the Governor (and Treasury and the Minister of Finance) are so keen on today.

As for the rest, sometimes financial stresses can exacerbate unemployment and the like,  but the financial crises typically arise and deepen in the context of common events or shocks that lead to both: people default on their residential mortgages when they’ve lost their jobs and house prices have fallen, but those events don’t occur in a vacuum.  And anyone (and Governor) who wants to suggest that mental health crises and a decline in social cohesion can be substantially prevented by higher levels of bank capital is either dreaming, or just making up stuff that sounds good on a first lay read.

The Governor ends with this sentence

If we believe we can tolerate bank system failures more frequently than once-every-200 years, then this must be an explicit decision made with full understanding of the consequences.

As if his, finger in the dark, once-every-200 years is now the benchmark, and if not adopted we face serious consequences.   Let’s see the evidence and analysis first.  Including recognition that systemic banking crises don’t just happen because of larger than usual random shocks –  the isimplest scenario in which higher capital requirements “work” –  but mostly from quite rare and infrequent bursts of craziness, not caused by banks in isolation, but by some combination of banks and (widely spread) borrowers, often precipitated by some ill-judged or ill-managed policy intervention chosen by a government.   Higher capital ratios just aren’t much protection against the gross misallocations that arise in the process –  in which much of any waste/loss is already in train (masked by the boom times) before any financial institution runs into trouble (the current Chinese situation is yet another example).

Perhaps as importantly, under the current (deeply flawed) Reserve Bank Act the choice about capital is one the Governor is empowered to make.  But his deputy, responsible for financial stability functions, had some comments to make on this point in a recent speech (emphasis added)

And Phase 2 of the Government’s review is an opportunity for all New Zealanders to consider the Reserve Bank’s mandate, its powers, governance and independence. The capital review gives us all an opportunity to think again about our risk tolerance – how safe we want our banking system to be; how we balance soundness and efficiency; what gains we can make, both in terms of financial stability and output; and how we allocate private and social costs.

It may be that the legislation underpinning our mandate can be enhanced, for example, by formal guidance from government or another governance body, on the level of risk of a financial crisis that society is willing to tolerate.

These are choices that should be made by politicians, who are accountable to us, not by a single unelected and largely unaccountable (certainly to citizenry) official.  We need officials and experts to offer analysis and advice, not to be able to impose their personal ideological perspectives or pet peeves on the entire economy and society.

We must hope that the forthcoming consultative document is a serious well-considered and well-documented piece of analysis, and that having issued it the Governor will be open to serious consideration of alternative perspectives.  But what was released last Friday –  700 words of unsupported jottings –  wasn’t promising.

(I should add that I have shifted my view on bank capital somewhat over the years, partly I suspect as a result of no longer being inside the Bank. It is somewhat surprising how –  for all one knows it in theory –  things look different depending on where one happens to be sitting.  But my big concern at present is not that it would necessarily wrong to raise required bank capital, but that the standard of argumentation from a immensely powerful public official seems –  for now – so threadbare.)

Implicit admissions and bids for resources

The Reserve Bank’s Financial Stability Report was released earlier this morning.  The headline, of course, was the easing in the loan to value restrictions on mortgage lending, although perhaps what should get more attention was the Governor’s suggestion that the avowedly “temporary” restrictions” will be in place for at least “the next few years”.     There was no good case for them –  putting a bureaucrat between willing borrowers and willing lenders – in the first place, and there is no good case for having them in place now.  Other than, of course, the interest that isn’t the public interest at all –  more discretionary power for an unelected unaccountable public official.

(Given the Bank’s repeated unease about dairy debt, it has also never been clear to me why LVR limits were appropriate for people buying houses but not for people buying farms. I used to raise the point while I was still at the Bank, and have never heard a satisfactory or persuasive response.)

Two other small things in the press release warrant just brief mention for now:

The first was this

Our preliminary view is that higher capital requirements are necessary, so that the banking system can be sufficiently resilient whilst remaining efficient. We will release a final consultation paper on bank capital requirements in December.

Time will tell how persuasive their case is, but given the robustness of the banking system in the face of previous demanding stress tests, the marginal benefits (in terms of crisis probability reduction) for an additional dollar of required capital must now be pretty small.

And the second was this

Aside from CBL, the insurance sector as a whole is meeting its minimum capital requirements. However, capital strength has declined and a number of insurers are operating with small buffers. The insurance industry must ensure it has sufficient capital to maintain solvency in all business conditions.

That is quite a shot across the bows of the sector, but it is worth remembering that when the solvency standards were set up the Reserve Bank consciously chose not to require insurers to hold sufficient capital to remain solvent in all circumstances. I vividly recall the day I asked, at the internal Financial System Oversight Committee, whether the solvency standards were demanding enough that they would have prevented the AMI collapse, and was told no.

But two other things caught my eye in the full document.

The first was that the Bank no longer seems to be claiming that LVR controls –  coming between borrowers and lenders for five years now –  have done anything to improve the soundness of the financial system (while they have inevitably impaired the efficiency of the system).  Those are the statutory goals the Bank is required to use its powers towards, and yet in the document today we find this (in the cartoon summary at the front):

The restrictions have reduced the number of borrowers who would be forced to sell their houses or significantly reduce spending if they ran into financial problems.

But, even if true, that is not the same –  at all – as improving the soundness of the financial system. It is about “nanny knows best” customer protection, which is no part of the Bank’s mandate.   You can’t be forced to sell your house if the Bank’s action prevented you from getting into one in the first place.

And here is the claim from the body of the document

The Reserve Bank’s LVR restrictions have leaned against the build-up in risks from high household debt by increasing the amount of equity borrowers have in their homes. The restrictions have seen the proportion of outstanding mortgage debt to households with loans larger than 80 percent of the value of their houses fall from over 20 percent in 2013 to under 7 percent. This extra equity provides households with more room to avoid cutting consumption or defaulting on their loans if economic conditions deteriorate or if interest rates rise.

Nannying again, not (apparently) focused on the soundness of the financial system.  As a reminder, the two diverge because (a) even if LVR controls modestly reduced housing lending risks, we never get a good sense of what other risks banks have taken on to maintain profits, and (b) because less risky lending means banks need to hold less capital.  Capital relative to (properly assessed) risk-weighted assets is the key issue when it comes to solvency.

From the text there is no way of telling whether the Bank’s focus has really changed or just the marketing. But marketing –  from a powerful public agency – should be aligned with, and disciplined by, mandate.

And then there is climate change.  In the Governor’s press release there was this

In the medium-term, an industry response to a variety of climate change-related challenges appears likely, requiring investment.

Which is pretty cryptic, perhaps even empty.

But in the full document there is a two page spread on “The impact of climate change on New Zealand’s financial system”.   There is lots of text, and very little substance.  It smacks of the Governor bidding for relevance – signalling to his buddies on the (political and business) left –  and involvement in the wider whole of government programme, and perhaps worse, it looks like a bid for more budgetary resources (a case we know the Bank has been making) or amended legislation to do things like

The Reserve Bank is developing its own climate change strategy. The strategy focuses on ensuring that climate risks are appropriately incorporated within the Reserve Bank’s mandate. The Reserve Bank also stands ready to collaborate with industry and government to help position New Zealand for the challenges ahead.

This for a body with two city offices, and a balance sheet full mostly of exposures to New Zealand government debt and overseas government debt.

The text burbles on about possible risks, but it all adds up to very little.     There are numerous risks banks and borrowers face every decade, every century.  Relative prices change, trade protection changes, external markets change, exchange rates change, technology changes, economies cycle, land use law changes.  Oh, and the climate changes.

If one looks at the structure of New Zealand bank (or insurer balance sheets) it just isn’t credible that climate change poses a significant risk to the soundness of the New Zealand financial system (that pesky law again).   Some individuals are likely to face losses from actual and prospective sea-level rises, but banks (and insurers) typically have diversified national portfolios.   People can’t have mortgage debt without insurance, and so the insurers are likely to be constraining people first.   Much the same surely goes for the rural sector?   Sure, adding agriculture into the ETS at the sort of carbon price some zealots have called for would be pretty detrimental to the economics of a dairy debt portfolio, but then freeing up the urban land market probably wouldn’t be great for residential mortgage portfolios, and we don’t see double-page spreads from the Reserve Bank on that issue, or the Governor trying to play himself into some more central role in that area.     It smacks of politics –  signalling the Governor’s green credentials –  more than anything legitimately tied to financial system soundness.

But then we probably should not be surprised. The Governor sells himself as head of a tree god (fortunately there was none of that stuff in today’s document), and gives speeches on climate change, but eight months into his term still hasn’t managed to give a speech on either of his main areas of statutory responsibility (monetary policy or financial supervision/regulation/stability).


LVR restrictions: towards the FSR

The Reserve Bank’s latest Financial Stability Report is due out on Wednesday.  Perhaps we will see some further articulation of the Governor’s strange vision of the Bank as a tree god, but I guess the main interest will be in what, if anything, the Governor does with the loan to value controls rushed into place, and then frequently amended, by his predecessor a few years ago.  It is as well to recall that although legislation is going through Parliament at present that will, at least on paper, modestly weaken the Governor’s personal power over monetary policy, in respect of banking regulation his statutory powers remain untrammelled, and unchannelled.  There are few legal constraints on what he  –  an unelected official whose appointment was controlled by unelected and unaccountable academics and company directors –  can do.

Market economists are, understandably enough, focused on the narrow question of whether there will be any changes to the rules announced this week. You can read a summary of their views here.   I remain less interested in that (forecasting) issue than in the cases for and against having such controls in the first place.   They are a new thing: we never had some legal restrictions in the bad old days of a heavily regulated financial system prior to 1984.  But, like weeds or wilding pines, once regulatory controls get in place people come to treat them as normal, the only debate tends to occur around the edges, and it takes huge effort to do something serious about fixing the problem.  Years ago, when the LVR restrictions were first introduced, we were assured they would be temporary (I was still inside the Bank at the time, and as far as I could see senior people genuinely believed it) but now the very idea that willing lenders and willing borrowers should be free to contract on mutually agreeable terms seems to be becoming lost.

The Herald’s economics columnist Brian Fallow used his column last Friday to argue to “Keep the brakes on houses”.  I can’t see the column on line, but the gist of his article is that house prices are high and household debt is high and that unless that combination changes the Reserve Bank shouldn’t think of lifting the LVR controls.  It doesn’t matter that stress tests repeatedly show that banks can cope with big falls in house prices and even big rises in the unemployment rate.  It doesn’t matter that our banks came through the last serious recession –  when household debt to GDP was about as high as it is now –  unscathed. It doesn’t matter that high house and land prices are mostly a phenomenon of the artificial scarcity created by land use restrictions (with high construction costs into the mix).  It doesn’t even seem to matter than there is no evidence that the LVR controls have made banks safer (banks with fewer individual risky loans also need to hold less capital) or the economy more stable.  It doesn’t seem to matter that the LVR controls have acted to favour established (cashed-up) buyers over new entrants to the housing market.  No, even though there is no threat to financial stability, and everyone recognises that LVR limits impede the efficient functioning of financial markets  (and those are the only two criteria the law allows the Bank to act under), the call is simply to leave the controls in place.

It was a bit like people in earlier decades who opposed removing import licenses or exchange controls because of the “foreign exchange constraint” (imports might increase if we took the controls off): papering over symptoms rather than tackling causes is rarely a sensible approach to policy.  Sadly, this government, like its predecessor, seems to be doing almost nothing to fix the underlying problem (and when I heard the UDA announcement over the weekend cited as something that had “worked well” in the UK and Australia one was reminded a new of just how obscene house prices in the UK and Australia remain).  But if the government isn’t doing anything serious, they will no doubt be grateful for the cover the Reserve Bank provides, claiming that somehow it is “doing its bit”, when it has no responsibility (there is “our bit”) for the fundamental problem.

But, of course, with no evidence whatever, the Governor is convinced that he knows best, the banks and markets are too “short-sighted” and so no doubt the controls will remain.  If the Governor is really so convinced he should at least really go to the effort of persuading the Minister of Finance to agree to extend the restrictions to other non-bank lenders.  The LVR controls only apply to banks because they are the only lenders the Reserve Bank Governor himself can order round in this way –  restrictions on other non-bank deposit-takers require the agreement of the Minister of Finance.  We have been fortunate in the last few years that there has been less disintermediation of mortgage business to non-bank lenders than most (including Reserve Bank staff doing the evaluation) had expected.   But if we learned anything from the decades of heavy controls prior to 1984, over time risk-taking will gravitate to institutions where it can occur.  Putting in place a competitively-neutral regulatory framework (treating banks and non-banks similarly) was a huge step forward in the 1980s, and it is unfortunate that the Reserve Bank now treats the same risks differently depending on whether an institution wears a “bank” or “non-bank” label.

At the press conference a few weeks ago for the Monetary Policy Statement, the Governor and his deputy (once a fairly market-oriented economist) indicated that in the forthcoming Financial Stability Review the Bank was planning to outline a more disciplined framework or road-map for assessing when, and whether, adjustments should be made to the LVR controls.  In principle, that sounds sensible and welcome. In principle, it should involve the Bank setting out some markers against which they can be held to account when the next decisions/reviews come round.  Whether it is so in practice only time will tell, but I was disconcerted when I heard them talking of this new framework as something similar to what they have for OCR decisions.

While we have a Reserve Bank there have to be regular monetary policy decisions.  That isn’t so for LVR restrictions, and it would be unfortunate if the initiative the Bank has foreshadowed was also about entrenching LVR controls as a permanent feature of New Zealand’s financial system.  Capital and liquidity requirements, backed by regular and robust stress tests, should remain the heart of our banking regulatory framework, rather than having bureaucrats reach into private businesses –  well-run over decades –  and tell them who they can and can’t lend to, or who they can and can’t employ.  But bureaucrats have incentives to build up their bureau and are typically reluctant to give up powers they’ve once got their hands on.  They are just human, and in their shoes you and I might face similar temptations.  We need banking regulation and supervision –  mostly, in my view, because politicians will bail out large banks in crises and everyone knowing that efforts need to be made to limit the risks –  but its appropriate place is distinctly limited, accountable, and kept in fully in check.   Instead, those paid to hold the Bank to account –  ministers, the Board, FEC –  mostly just accommodate the regulators, at times even egging them on.

On a totally different matter, for anyone interested in a some snippets of New Zealand economic history, you might want to try the latest Newsroom/Radio New Zealand Two Cents’ Worth podcast, which was built around the odd coincidence that –  if you look at the data a bit loosely and from the right angle – for several decades in a row, years ending in 8 had also seen a New Zealand recession.  I did an interview with Bernard Hickey for the podcast, in which he had me run quickly through aspects of the New Zealand economy in each “8” year since 1918.   As I noted to Bernard, there is now a rather large hole in the market for a up-to-date economic history of New Zealand (the last full one appeared in about 1985 and much has been done, much has happened, since then).  Brian Easton’s long-awaited history of New Zealand from an economics perspective – his framing –  is still awaited.

More thoughts on financial crises and economic performance

In my post yesterday, focused specifically on Geoff Bascand’s speech on financial stabilty, financial crises etc, I used this chart

crisis costs

to, again, raise questions about just how much of the poor economic performance over the last decade or so can really be ascribed specifically to the financial crisis (bank failures, large loan losses etc).  After all, the US was the epicentre of the crisis, and my other group of countries (long-established advanced countries, also with floating exchange rates –  Australia, Canada, New Zealand, Norway, Israel, and Japan)  didn’t have domestic financial crises.

I’d been playing around with that data with a view to writing a post about an article in the latest issue of Foreign Affairs, The Crisis Next Time: What We Should Have Learned from 2008″, by Carmen and Vincent Reinhart (she an academic researcher, and he a senior market economist and formerly a senior Fed official).    The Foreign Affairs website is having open access this month, so the link should work for anyone wanting to read the (accessible and not overly long) article itself.

I thought the article was a bit of a mixed bag (and this post ends up only partly being about the article).  Carmen Reinhart, in particular, has been at the forefront of efforts to remind that recessions associated with financial crises are often more severe than other recessions.  That is a useful reminder, but hardly surprising.  Mild recessions tend not to generate many loan losses, and even if the banking system wasn’t rock solid in the first place, nothing too serious is likely to follow.  But if resources have been severely misallocated in the first place, supported by ample new credit, then when the correction occurs –  and views about what is profitable have to be revised –  it isn’t surprising that the associated recession can be deep and the financial system can come under stress.  In New Zealand, for example, it wasn’t the financial system crisis (failure of DFC, repeated near-failures of the BNZ) that made the 1991 recession so serious; rather than pressures on the financial system were part of the same aftermath of excess –  over-inflated expectations – that the entire economy was caught up in, combined with some serious efforts to break the back of high trend rates of inflation.

As the Reinharts point out, the problems can then be particularly severe in a country that has few or no macro policy levers left open too it –  a fixed exchange rate or a common currency, tied to the fortunes of a group that may not share the particular problems you did (thus, for example, Ireland in a euro-area in which Germany is the largest economy).  Adjustment can be a lot slower without the ability to adjust the nominal interest and exchange rates.  Perhaps more than the authors, I’m a sceptic on the euro.

For my purposes, there is a convenient couple of sentences in the Reinhart article

Financial crises do so much economic damage for a simple reason: they destroy a lot of wealth very fast. Typically, crises start when the value of one kind of asset begins to fall and pulls others down with it. The original asset can be almost anything, as long as it plays a large role in the wider economy: tulips in seventeenth-century Holland, stocks in New York in 1929, land in Tokyo in 1989, houses in the United States in 2007. 

It usefully highlights a key difference between, say, the US (or Ireland or Iceland) late last decade, and the experiences of the group of non-crisis floating exchange rate countries whose experience is reflected in that first chart above.   Stock markets in those latter countries took a short-term hit, of course, but there was no sustained loss of (perceived) wealth akin to what happened in the crisis countries.

It isn’t entirely clear from the article how much the authors want to focus mostly on the depth of the initial recession and how much on the disappointing economic outcomes in many countries over the last decade.  But both are mentioned, and there seems to be a tone that conflates the two in a way that I’m not surely is overly helpful (given the goal of learning lessons that can help better prepare us for future severe adverse events).  There also seems to be a very strong focus on the demand side, and none at all on the supply side (no mention at all of productivity growth).

And yet, if we look across the OECD as a whole, the unemployment rate was right back down to where it had been in 2007.  If (and there is) a disappointment about the last decade as a whole, it can’t be now about excess labour supply (unemployed workers) –  slow as the unemployment rate was to come down, it did eventually.  As it happens, the unemployment rate in the US (epicentre of the crisis) is now lower than in the median of my non-crisis floating exchange rate group –  which wasn’t the case in the years running up to 2008.

I have plenty of criticisms for the way many central banks (including our own) handled the years after the 2008/09 crisis and recession.  In some cases, actually tightening when it wasn’t necessary or appropriate, and often a hankering for some sort of return to “normal” interest rates (that may have prevailed in the previous couple of decades) when as has become increasingly apparent something about what is “normal” has changed.  Throw in the lack of any pro-activity in addressing the existence of the near-zero lower bound on nominal interest rates (itself arising from regulatory and legislative choices), and it is clear that more could –  and should –  have been done in many countries.

But even if such changes (in macro policy) had been made, the differences in economic outcomes would probably have been at the margin:  helpful (eg in a New Zealand context, getting core inflation back to 2 per cent, and getting unemployment down to the NAIRU perhaps two or three years earlier), but it is unlikely that it would have made much difference to productivity growth, or indeed to levels of real GDP per capita today.

In yesterday’s post, I showed a chart comparing labour productivity growth trends in the US (epicentre of the financial crisis) and in the group of non-crisis floating exchange rate advanced economies.  But what about multi-factor productivity?

The OECD only has MFP data for a subset of member countries.  Of my sample of non-crisis advanced countries, they don’t have data for Norway and Israel.  But here is the comparison for the US and the group of four non-crisis advanced countries, all normalised to 2007.

MFP crisis.png

In both cases –  although perhaps more starkly so for the non-crisis countries –  it is clear that the slowdown in productivity growth was underway well before the recession (and crisis).  The financial crisis (centred in the US) cannot be to blame for something that is (a) apparent across crisis and non-crisis countries (especially when the non-crisis countries are less productive than the US to start with), and (b) when the phenomenon got underway before the crisis or recession did.

(The Conference Board Total Economy database does have MFP estimates for my full group of non-crisis countries.   They use a different model to estimate MFP, but the same two key observations hold in their data: the slowdown was apparent in both lots of countries well before the crisis/recession, and (if anything) the US has done better than the non-crisis group both before and since its crisis.)

But what about some of the euro-area countries you ask?  And the Reinharts themselves rightly point out how poor the economic performance of Italy (and Greece) has been.  The OECD doesn’t have MFP estimates for Greece, but here are the estimates for three other embattled euro-area countries: Portugal, Spain, and Italy.

MFP crisis 2

All three countries have been in deep trouble for a long time now –  the estimated level of MFP peaking around 2000.   On this score, the trends don’t look materially different over the last decade than over the years leading up to 2007.    Whatever the cause of their problems with productivity, it can’t have been the financial crises these countries went through.

And perhaps nor would you expect it.  Readers might recall a wrenching financial crisis that Korea went through in 1998.   And here is the OECD estimate of multi-factor productivity for Korea.

mfp crisis 3

You can see the 1998 crisis/recession in the data, but as a short-term blip.  In the decade after the crisis, Korea productivity growth kept on at much the same rate experienced in the decade prior to that crisis –  before (presumably) joining in the global slowdown this decade.  (That had also been the experience of the United States in earlier crisis episodes –  estimates suggest that the 1930s, for all its problems (around demand shortfalls) was a period of strong MFP growth.)

There is lots to learn from the searing experience of crisis, recession, and slow growth in the advanced world over the last decade or more.   But I still reckon there needs to be a much more careful unpicking of the different strands of the story than central bankers –  who tend to see the world through money and finance lenses, and who are often keen to champion their future role –  are prone to.  To me, the cross-country evidence just doesn’t square with a hypothesis in which the financial crisis itself plays any large part in the sustained disappointing performance of so many countries over what is now such a long time.

Central bankers meanwhile might be better off rethinking the merits of arrangements like the euro, or of the continued passivity around the near-zero lower bound, both of which look as though they have the potential for causing very major problems the next time there is a serious economic downturn.

We need better foundations for financial stability policy

Adrian Orr is now 7.5 months into his term as Governor and we still haven’t had an on-the-record speech from him about either main strand of his responsibilities: monetary policy or financial supervision and regulation.  Is he just not engaged on these issues?

But yesterday, his deputy Geoff Bascand delivered  –  in Australia –  a substantive speech on financial stability issues.  There were a few good elements in the speech.  For example, I was pleased to see this in the conclusion

The capital review gives us all an opportunity to think again about our risk tolerance – how safe we want our banking system to be; how we balance soundness and efficiency; what gains we can make, both in terms of financial stability and output; and how we allocate private and social costs.

It may be that the legislation underpinning our mandate can be enhanced, for example, by formal guidance from government or another governance body, on the level of risk of a financial crisis that society is willing to tolerate.

At present, the legislation is drafted so broadly and loosely that a single unelected and unaccountable official gets to make any such choices.  He (as it typically is) gets to make choices in a pretty much unconstrained way and we (including our elected political leaders) just have to live with the consequences.    Whether the sort of formal guidance Geoff refers to in that second paragraph is (meaningfully) feasible is open to question, but we need to improve on the current situation.  If such guidance isn’t feasible –  if society can’t write down its preference and give them as a mandate for the technocrats –  the big decisions around banking supervision policy frameworks (as distinct from the application of them to individual institutions) should be made by elected politicians (the Minister of Finance).

But, sadly, most of the speech just wasn’t that good.  It had plenty of politicially popular lines, and there was even the obligatory reference to the Reserve Bank as a tree god.  On climate change we had this

Climate change presents significant financial stability risks both through the direct implications of physical events for insurers, farmers and households, the indirect effects on insurance availability and property values, and through the potential social and economic disruption it promises.

We are working on developing a climate change strategy, which will be informed by discussions with banks and insurers in due course. Our role as a regulator is to try to ensure that financial institutions are adequately managing these risks, even though the horizon for their realisation could be decades away.

Given that the best evidence for New Zealand is that projected increases in global temperature are probably neutral and at best slightly positive for New Zealand in economic terms, and that all sorts of relative price changes occur every year changing the economics of all manner of businesses banks might have lent on, all this should amount to nothing.  But we know the Governor is a zealot –  why, he bets billions of dollars of your money on particular views of the economics of climate change, while so obscuring the choice there is no effective accountability –  so no doubt there will be pages and pages of bureaucratic bumpf from an agency with no expertise in the issue (or mandate), simply adding to compliance costs (especially for small institutions).

There was a rather lame attempt to defend the Bank’s involvement in the bank conduct review.  I noticed that the Governor had a bit of spat with ACT MP David Seymour at FEC last week on just this issue, which ended with the Governor (to whom any concept of deference or politeness seems unknown) responding as follows

When Seymour persisted, Orr simply said: “I am right, you are wrong”.

My own take is that they are probably both right.  Seymour is right on the fundamental point –  the bank conduct review was about politics and perhaps about Orr advancing his standing, not about financial soundness and efficiency (the Bank’s statutory mandate).  And if Orr is correct –  about the law giving him scope to do this –  it is only because the legislation was written –  guided by Bank officials – far too broadly in the first place.

But what bothers me rather more is the Bank’s weak understanding of the nature of financial crises, systemic risks, and so on.  These are concerns I’ve raised over several years in various contexts, including the cases the Bank has made for LVR restrictions and the (longed-for) debt to income restrictions.

For example, they continue to claim that

Household sector indebtedness represents the New Zealand financial system’s single largest vulnerability.

Yes, household debt is the largest component of financial system assets, but that is a quite different proposition.  As their stress tests have repeatedly shown, banks’ housing portfolios are constructed in a sufficiently cautious way that even very large adverse shocks (rising unemployment and falling house prices) wouldn’t threaten the soundness of the banks.   They run this cross-country chart of credit to households as a share of GDP.


Yes, there is a lot more household credit than there was. That is the inevitable consequence of things like land-use restrictions than make urban land artificially scarce (and highly-priced).  And in New Zealand’s case, household debt to GDP is still a touch lower than it was going into the last recession (and at that time the servicing burden was also much heavier).  Despite all the angst, bank housing portfolios came through that severe recession unscathed –  as they did in Australia, Canada, and the UK.

But perhaps my biggest problem with the speech is a combination of three things:

  • the attempt to suggest that the system is very fragile –  at least without wise bureaucrats –  and that crises are always just around the corner, coming for us,
  • the continued failure to pay attention to the experiences of countries that had significant asset and credit booms and didn’t have a domestic financial crises, and
  • the inexcusable failure – in a central bank –  to distinguish between countries with floating exchange rates (which greatly assist adjustment in the face of shocks) and those without.

In combination, the Reserve Bank leads us towards quite misleading conclusions about the economic costs of financial crisis.  By overstating those costs –  hugely overstating –  they seek to strengthen their own position (and our respect of them) as regulators; the people who will do everything to keep us safe. (As commonly, one never sees mentioned in the speech that in all the financial crises they like to cite, there were in fact banking regulators who no doubt thought they were doing their job well.)

Of my first bullet, they say

First, why does financial stability matter? The answer is that bank crises are frequent and bank crises hurt.

Since the mid-1970s there have been over 140 banking crises around the world.

and (without any backing for this claim)

Serious incidents (that could have led to a crisis) are more common than people realise.

Yes, there have been lots of crisis, although since (depending on your definition) there are getting on for 200 countries in the world, even the number the Bank cites is less than one crisis per country over 45 years.

But there haven’t been many at all in stable, well-managed, floating exchange rate countries.  And in countries like ours –  for example, New Zealand, Australia, Canada, Norway –  the only financial crises in 100 years have related to the period just after liberalisation when everyone was just getting grips with what a market financial system meant (and when, for that matter, regulators also didn’t cover themselves with glory).    Of course, well-run banking systems can run into trouble, but since it is New Zealand that our Reserve Bank is supposedly focused on one might expect some grounding in the Australasian experience.   That experience just doesn’t suggest danger (massive credit losses) lurks continually.

The Reserve Bank has long been keen on citing the experience of Ireland as somehow relevant to New Zealand.  It pops up again in this speech

The consequences in terms of employment are also severe. After the GFC, Ireland’s unemployment rate rose from 4.6 percent in 2006 to 15 per cent in 2012

And yet –  prosperity and geography aside –  what is the biggest relevant difference between New Zealand and Ireland?   We get to set our own interest rates, and our exchange rate can adjust freely, while Irish monetary policy is set in Frankfurt for the entire euro-area, and they have no nominal exchange rate to adjust.  The Reserve Bank knows very well that floating exchange rate exist in large part because they provide greater leeway to cope with severe adverse economic or financial shocks.  Thus it was from the beginning –  at the time of the Great Depression –  and is now too.    I did post a few years ago –  which I can’t now see –  documenting that no floating exchange rate advanced country has ever experienced an increase in its unemployment rate of the magnitude Ireland put itself through.  I could commend to the Reserve Bank the experience of Iceland (which went through a financial crisis which, in many respects, was even nastier than Ireland’s, and yet had only a fairly moderate increase in its unemployment rate).

And then there is the hoary old chestnut about just how expensive financial crises supposedly are.  Here is Bascand

Since the mid-1970s there have been over 140 banking crises around the world. And they have had large costs for the affected economies and societies.

On average a bank crisis costs a country 23% of its GDP, while public debt increases by around 12 percent.3 The amounts are higher for advanced economies.

That footnote records that the numbers are calculated as deviations of actual GDP from its (pre-crisis) trend.

They sound like scary numbers, and if true (in some meaningful sense) they might even be so  (although even if a crisis happens every 20 years, a loss of 23 per cent of one year’s GDP is roughly a loss of 1 per cent of the total GDP over that full period).  But they aren’t meaningful, on a number of accounts.

First, the calculations (implicitly) assume that any deviation from the pre-crisis trend is a result of the crisis itself –  and not, for example, the misallocation of real resources that might well have occurred even if the financial system had stayed sound.  At best, these numbers conflate the two effects.

Relatedly, the estimate ignore things that might have getting underway in the year or two prior to the crisis.  Thus, as I’ve shown before, productivity growth in the United States had already begun to slow very markedly a couple of years before the crisis hit.


A small amount of that might make its way into the pre-crisis trend measures, but most of it won’t.

And thirdly, the Bank –  and many of their peers among other keen regulators –  makes no attempt to compare the experiences of countries that went through serious financial crisis and those that did not.   US economic performance over the last decade has been underwhelming to say the least.  The US was at the epicentre of the 2008/09 financial crises.  But it is simply a step far too far to conclude that the extent to which the US has done less well than in the previous decade is the measure of the cost of the financial crisis, especially if other countries that didn’t have a crisis also did less well than they had done the previous decade or so.

I’ve touched on this issue before, including in this post last year.   Of course, finding good comparators isn’t just a matter of a random into the OECD bag of countries.  For a start, as I’ve already noted (and as the Reserve Bank knows), a fixed exchange rate tends to exacerbate the severity of any shock.  The United States –  epicentre of the financial crisis –  is a floating exchange rate country.   Some floating exchange rate countries –  notably the UK and Switzerland –  were caught up in the 2008/09 crisis primarily because of the exposure of their internationalised banking sector to the US and its housing debt instrument (rather than because of domestic credit exposures).  But there are six well-established floating exchange rate advanced countries that didn’t have a serious domestic financial crisis at all in 2008/09:  New Zealand, Australia, Canada, Norway, Israel, and Japan.

Here is how the US experience, on real GDP per capita, compares with the median of those non-crisis floating exchange rate advanced economies

crisis costs

The US experience was a little worse than that of the median of this group of countries, but the differences are small, and there is a lot of variability in the experience of the non-crisis countries (since 2007 Israel has done much better than the US, while Norway has done much worse).   And as I noted in the earlier post, the comparison still tends to exaggerate any contribution of financial crises themselves, as the US had less fiscal leeway than all the other floaters except Japan, and the US had less monetary policy leeway (running into the lower bound) than New Zealand, Australia, and Norway.

That’s GDP per capita.  But what productivity?  Quite a lot of the arguments about the cost of financial crises attempt to build a story about persistent dampening effects on innovation, risk-taking etc, reflected in the productivity numbers.  Here is the chart, showing the same comparison countries, for real GDP per hour worked (OECD data).

crisis costs 2

Perhaps this chart is a bit more favourable to the story, depending on how you read it. Over the whole period –  pre and post crisis –  the US managed faster labour productivity than the median of the six non-crisis countries.  But perhaps the slowdown in productivity growth is a bit more in the US than the others (even if, as the earlier chart showed more clearly) the slowdown pre-dated the crisis?  Then again, the level of labour productivity in the US is higher than in all but one (Norway) of my non-crisis collection of countries, so if there was a global productivity growth slowdown (for whatever reason) you might be expect the US to be hit more visibly than the other countries (that sitll had catch-up and convergence opportunities).   Even among the non-crisis countries, there is considerably divergence –  since 2007 Australia has had the strongest productivity growth and Norway the weakest.  (Remarkably, Iceland –  savage financial crisis and all –  has had faster labour productivity growth than all these countries.)

I’m not wanting to suggest that recessions and financial crises don’t have costs.  At an individual level almost inevitably they do, and at a national level recessions are rarely pleasant or welcome (that’s why we have active monetary policy).  But we deserve much more searching analysis from our central banks and financial regulators (and those holding them to account, including national Treasurys) when they bidding to persuade us to entrust them with so much power, and  the deference due to people who make so much difference (so they claim).

A good starting point remains this very long-term chart (due originally to Nobel laureate Robert Lucas)


As I noted in a long-ago post

It is a quite simple chart of real per capita GDP for the United States, back as far as 1870.  These are Angus Maddison’s estimates, the most widely used set of (estimated) historical data, and as Maddison died a few years ago they only come as far forward as 2008.  The simple observation is that a linear trend drawn through this series captures almost all of what is going on.  More than perhaps any other country for which there are reasonable estimates, the United States has managed pretty steady long-term average growth rates over almost 140 years.  And yet, this was a country that experienced numerous financial crises in the first half the period.  Lists differ a little, but a reasonable list for the US would show crises in 1873, 1884, 1893, 1896, 1901, 1907, perhaps 1914, and 1929-33.  There were far more crises than any other advanced countries experienced.

And yet, there is no sign that they permanently impaired growth, or income.

If we are to have good financial stability policy, and confidence in it, it needs to be based on good searching robust and honest analysis, that recognises the puzzles and the ambiguities in the data, not the sort that rushes to support the conclusions policymakers have already settled on.

The Fed and Lehmans

On the day of the US mid-term elections it seems appropriate to have a US topic.

I read a lot of books each year.  Many of them provide a fresh angle on some or other issue I’m interested in, but few lead me directly to change my mind.  Professor Laurence Ball’s The Fed and Lehman Brothers is one of the exceptions.   I wasn’t pre-disposed to expect much from Ball (a professor of economics at Johns Hopkins university): my impressions of him were formed by his visit to New Zealand 20 years ago when, as Reserve Bank professorial fellow at Victoria University, he somewhat embarrassed his hosts by suggesting that the conduct of key elements of fiscal policy should be handed over to independent technocrats.  Interesting idea I suppose, but given that the point of spending public money on the fellowship had been to buttress public support for an independent Reserve Bank, it didn’t really help, especially in an election year with Winston Peters in the ascendant.

But the new book looked intriguing. As it turned out, it was much more than that, and I’d go as far as to call it a “must-read” for any serious student of the 2008/09 financial crisis.

It is a very careful and detailed study focused largely on one question: could the US authorities have lawfully prevented the failure of Lehmans that fateful weekend in September 2008 if they had wanted to?   Key decisionmakers have claimed, at the time and subsequently, that there were no lawful options open to the Fed (Bernanke, for example, is quite explicit in his claim that the authorities could only have intervened in breach of the law).  Ball shows, pretty conclusively, that such claims are simply wrong.  The decision not to provide liquidity support to the Lehmans group was just that, a choice.  And he goes on to illustrate that although in law any decision to have provided liquidity support (or not) rested solely with the Board of Governors of the Federal Reserve, in fact the key player was the US Secretary to the Treasury, Hank Paulson, with the Fed apparently deferring to his preferences.

Under the Federal Reserve Act as it stood in 2008, the Fed could lend to non-banks (as Lehmans then was, and as Bear Stearns had been) only in “unusual and exigent circumstances”.  Most commentators will agree that in September 2008 –  a year into an unfolding financial crisis, shortly after the US government had intervened to support the mortgage agencies –  that particular strand of the legal test could readily have been passed, in respect of a major investment bank closely intertwined with the rest of the wholesale financial system in the US and abroad (Lehmans had major operations in London).  The other strand of the legal test was that any loans had to be “secured to the satisfaction of the Reserve Bank” making the loan.  There apparently wasn’t much (or any) case law on this provision, but it was generally accepted within the Fed that the Federal Reserve shouldn’t be lending if they weren’t pretty sure of getting their money back.

But what wasn’t in the statute was a requirement that the borrower itself still be solvent (positive net equity).   A financial institution’s directors would presumably have quite severe limits on their ability (or willingness to risk doing so) to trade while insolvent, but from the point of view of the Federal Reserve, considering providing lender of last resort liquidity support, the relevant issue wasn’t the solvency of the institution, but the adequacy of the specific collateral the Fed would receive to cover any loan.  Nonetheless, senior policymakers have since claimed that Lehmans was insolvent and that, in any case, there was insufficient good collateral to support a loan of the size that might have been required.    Ball challenges both claims.

He does so using an array of published material, including regulatory filings, bankruptcy examiners’ reports, and the report (and supporting documents) of the Financial Crisis Inquiry Commission.

On the solvency front, one issue Ball has to grapple with is that when Lehmans was placed in bankruptcy there proved to be a considerable shortfall in net assets: not just shareholders (who lost everything) but creditors lost significant sums (and some court cases are still unresolved).   But that is a quite different issue from whether there was positive net value in the business at the point where the decision not to provide liquidity support was being made.  Economists have long recognised the concept of “bankruptcy costs”, and Ball makes a pretty compelling case that the bankruptcy process itself resulted in significant transfers of value to other parties that would be unlikely to have occurred in a more orderly process (the three areas he singles out relate to the termination of derivatives contracts, the fire sale of subsidiaries, and the disruption of various investment projects (mainly in real estate) that Lehmans was party to.  But on a going concern basis Ball concludes his detailed analysis this way

…the best available evidence suggests that Lehman was on the border between solvency and insolvency based on realistic mark-to-market accounting, and it was probably solvent based on its assets’ fundamental values.

As noted earlier, the critical (legal) criterion wasn’t about institutional solvency, but about the specific collateral the Fed could have obtained.

You might have assumed –  in a hazy way I think I did –  that by the end Lehmans wouldn’t have had much decent collateral left.  Perhaps you assumed that if the Fed had lent, it would all have been “secured” on dodgy commercial real estate loans.  But, as Ball demonstrates, that view is quite wrong.    Lehmans had been funding a large proportion of its balance sheet (as was the norm then for investment banks) through repos using fairly high-quality securities (ones that Fed was happy to accept), and the run on Lehmans primarily took the form of counterparties not being willing to roll over this repo finance (itself an interesting phenomenon, given that repo contracts should have left any counterparty with a clean ownership of the collateral security in the event of bankruptcy). But to the extent the repos didn’t roll over –  and it was clear they wouldn’t have on the Monday morning without Fed support – Lehman would still have been left with the (good quality) securities.  It also had long-term funding on its balance sheet, which couldn’t go anywhere in the short-term.  Ball demonstrates that Lehman had sufficient volumes of good quality acceptable collateral that it could have secured a large enough Fed loan to have replaced all its short-term funding if necessary.   The numbers would have been large, but as Ball points out no larger than the amounts injected into AIG a few days later (for a risky equity stake), or lent to Morgan Stanley a short time later.

There is an important distinction to be made here.  The issue Ball is dealing with is not whether the US authorities should have taken over, and recapitalised, Lehmans.  His argument –  nested in the liquidity provisions of the Federal Reserve Act –  is that liquidity support could (lawfully) have been provided, and that had it been provided it would have opened the way to a less costly, less disruptive, resolution over the following months.  Perhaps it would have been possible to inject more private equity to the holding company and enable it to continue as a going concern.  But if not, the prospects for a takeover of the business would have been greater –  for example, a key obstacle to Barclays taking over Lehmans was the need for a shareholder vote which would have taken at least a month –  or it would have been possible to have sold subsidiaries –  including the valuable asset management subsidiary –  in a more orderly and competitive process.  At worst, a more orderly wind-down would have been facilitated.

One of the other things Ball documents is the work that had gone on inside the Fed over several months, right up to the fateful weekend, on possible liquidity support mechanisms for Lehmans.  It seems pretty clear that there was never a presumption inside the Fed that if a private buyer was not be found that Lehmans would simply be left to the tender mercies of the bankruptcy administrator.  (In fact, as he notes even when Lehmans was forced to file for bankruptcy, the Fed provided substantial liquidity support to keep the New York broker-dealer subsidiary open for several days until Barclays committed to purchase it.)

So why didn’t the Fed prove willing to provide liquidity support for the whole group?  Ball argues, pretty conclusively, that the key player here was Secretary to the Treasury, Hank Paulson.  In law, the Secretary to the Treasury (or anyone else in the Administration) had no role in such decisions.   And it is not as if, in the specifics of the time and system, Paulson had any greater political legitimacy than, say, Bernanke.  Both were appointed by (outgoing) President Bush, and both had been confirmed by the Senate.   Presumptively, Paulson was likely to be out of office in January 2009 no matter who won the election, while Bernanke had more of his term to run.  But, of course, the politics around Wall St “bailouts” had been turning increasingly nasty since the Bear Stearns intervention (where the Treasury had got involved, implicitly underwriting the Fed’s credit risk) and Paulson –  a strong personality –  was quite open that he didn’t want to be remembered by history as Mr Bailout.  Perhaps the distinction between well-collateralised liquidity support and (actual or implicit) equity support got bypassed in the heat of the moment.

But the other relevant aspect, given the political aversion to more “bailouts”, seems to have been a sense within the Fed that the pressures on Lehmans had been so well-foreshadowed, over months, that its failure wouldn’t prove that disruptive.  Key players now claim that that wasn’t their view –  Bernanke is on record claiming that he always knew it would be a “catastrophe” –   but Ball demonstrates that such claims are simply inconsistent with what the Fed was saying or doing at the time.  For example, the FOMC met two days after the Lehmans failure.  Had the Fed thought the Lehmans failure would prove “catastrophic”, or even just aggravating the severity of the recession, a cut to the Fed funds rate would surely have been in order.  There wasn’t one.  And the published records of the meeting show no sign of any heightened concern or anxiety about the financial system or spillover effects to the economy.  If that was the prevailing view at the top levels of the Fed, it makes more sense as to why central bankers would defer to political pressure not to have provided (liquidity) support for Lehmans.

Central bankers don’t emerge with much credit from Ball’s book.  Anyone can make mistakes in the heat of the moment –  even a large institution with a deep bench like the Federal Reserve –  but what is perhaps more troubling is the suggestion (which seems pretty convincing to me) that key players (Bernanke, Geithner and Paulson) had been spinning the situation in their memoirs, rather than confronting the specifics of the data and the law.  Perhaps I become a bit more sympathetic than I was to (former BOE Governor) Mervyn King’s choice to avoid memoirs, and a defence of his involvement, in his own post-crisis book.  Thank goodness then for the efforts of a careful, apparently dispassionate, academic like Ball.

Of course, to agree with Ball’s conclusion that the Fed could have provided liquidity support to Lehmans if it had wished to do so is not to immediately jump to the conclusion that they should have done so.   Although it isn’t the focus of his book, it is pretty clear that Ball thinks such support should have been given.

A counter-argument could have a number of strands:

  • first, Lehmans had been under pressure for months to raise additional outside equity, and had failed to do so.  Had they done so, even at deeply discounted prices, it is unlikely that the wholesale run would have developed as it did (and even had it done so, the politics of liquidity support might have been different),
  • second, had Lehmans been a bank supervised by the Fed it would probably not have been allowed to stay open even as long as it did without new capital.  In bankruptcy courts, the relevant test might be whether there are still positive net assets, but bank supervisors who are doing their job should have been intervening pretty strongly –  including using directive powers –  before any question arose as to whether net assets were still (perhaps barely) positive, and
  • third, there is still the unanswered question (which may never be satisfactorily resolved) as to just how much the Lehmans failure exacerbated the recession.  Counterfactual history is hard.   The consensus view at present is that the adverse effects were large, but if much of the disruption would have happened anyway –  even if Lehmans had been left limping for a couple of months on liquidity life-support –  the case for intervention is weaker than many would allow (and, for example, AIG’s plight was largely unrelated to the Lehmans failure).  After all, there is a salutary place for market discipline, including around the urgency of injecting new capital when dark clouds loom.

I was one of those who tended to welcome the decision not to “bail out” Lehmans (better still not to have intervened around Bear Stearns months earlier) but I probably haven’t distinguished clearly enough between liquidity and solvency support.   The latter option –  which wasn’t something the Fed could have done anyway –  isn’t the focus of this book, but Ball does make a pretty persuasive case around liquidity support, including based just on facts that were available at the time (on the aftermath, no one could be certain).

I could still mount a counter-argument based on the first couple of bullet points above.  Providing liquidity support in such circumstances would have sent a signal to boards and managers of other institutions that any urgency to raise new capital, at deep discounted prices, was less than it might have seemed.  On the information availabe at the time, that would have been unfortunate.   Then again, within days that whole argument was tossed out the window as the authorities rushed to respond to a deepening crisis.

But perhaps what finally gets me over the line in thinking the Fed made a mistake, in not lending and in deferring to Paulson (in a politicised time six weeks out from an election), is an assessment of the probabilities.  Perhaps the Lehmans failure really wasn’t that big a deal.  Perhaps the Fed at the time was justified in its view that a failure could be managed without too much spillover downside.  But operating in a world of heightened uncertainty, no one could really know.  There had to be a chance that simply allowing Lehmans to go into bankruptcy –  the largest bankruptcy in US history, all done in rush –  would prove very very disruptive and economically costly.  But if providing strongly-collateralised liquidity support, quite possibly at a high interest rate and with ample haircuts, could have alleviated that risk –  even if it was only a 10 per cent risk – it is hard not to conclude (even without the benefit of hindsight) that the central bank should have acted.  After all, lender of last resort provisions are put in statutes for a purpose –  and not just a decorative one.



Looking back to the deposit guarantee

12 October 2008 was a frantic day.  It was a Sunday, and I never work Sundays (well, two financial crises, one in Zambia, one in New Zealand, in 30+ years).  There was a call in the middle of our church service summoning all hands to the pump, to put in place a retail deposit guarantee scheme that day.   We did it.  My diary later that night records that we’d “delivered a brand spanking new not very good deposit guarantee scheme”, announced a few hours earlier.   It was a joint effort of the Reserve Bank and The Treasury.

I had recently taken up a secondment at The Treasury.  I’d been becoming increasingly uneasy about the New Zealand financial situation for some months (flicking through my copy of Alan Bollard’s book on the crisis I found wedged inside a copy of an email exchange he and I had had a month or so earlier about Lender of Last Resort options for sound finance companies, potentially caught up in contagious runs) but I hadn’t had any material involvement in the unfolding sequence of finance company failures.   But it was the escalating international financial crisis – this was four weeks after Lehmans, 3.5 weeks after the AIG bailout, two weeks after the US House of Representatives initially voted down TARP, and two weeks after the Irish government surprised everyone by announcing comprehensive deposit guarantees –  that really accelerated interest in the question of what, if anything, New Zealand should do, or might eventually be more or less compelled to do.    The initiative for some more pro-active planning came from The Treasury, but with some parallel impetus  –  including around guarantees – from the then Minister of Finance, Michael Cullen (who, a few days out from Labour’s campaign launch, was also looking for pre-election fiscal stimulus measures).

On Tuesday 7 October, there was a long meeting at the Reserve Bank, attended by both the Secretary to the Treasury, John Whitehead, and the Governor of the Reserve Bank.  My memory – and my contemporary diary impression – is that the Governor was considerably more focused on the managing the Minister’s political concerns than on any sort of first-best response.    But the outcome of that meeting was agreement to quickly work up a joint paper for the Minister which would not, at that stage, recommend introducing a deposit guarantee scheme, but which would outline the relevant issues and operational parameters, giving us something to work from if the situation worsened.

Which it quickly did, both on international markets, and with the political pressure, with the Prime Minister signalling that she wanted to be able to announce something about guarantees in her campaign launch that coming Sunday afternoon.

I and a handful of others on both sides of The Terrace scurried round for the next few days.  I see that in my diary I wondered what the best approach was: do nothing, allow some risk of the crisis engulfing us, and then pick up the pieces afterwards, or be more pro-active and take the guarantee route.  My conclusion –  and even today I wince at the parallel (but this was a late-at-night comment) – “I suspect that if the pressures really come on, the Irish approach is best”.   As relevant context, although much of the finance company sector was in solvency trouble (many had already failed) there were no serious concerns about the solvency of the banking system.   (Liquidity was, potentially, another issue.)

At Treasury we had recognised the importance of the Australian connection –  most of our banks being Australian-owned.     I’m not sure of the date, but we had taken the initiative –  at Deputy Secretary level –  of approaching the Australian Treasury to see if they were interested in doing some joint contigency planning around deposit guarantees, and had been told that the Treasurer had no interest in such guarantees and so our suggestion/offer was declined.

But even Australian authorities could look out the window and see that the global situation was deteriorating rapidly, and by late in the week that recognition was being passed back to authorities on this side of the Tasman.  Alan Bollard always kept in close contact with his RBA counterpart Glenn Stevens, and on the Friday my diary records (presumably told by some RBNZ person I was working with) “apparently Glenn S[tevens[ told Alan this afternoon that the RBA/authorities might fairly soon have to consider a blanket guarantee”.     In the flurry and uncertainty, one other senior RBNZ person –  still holding a senior position there –  told me that in his view nothing should be done here unless there were queues outside New Zealand banks.

Between a handful of people on the two sides of the street, we got a paper on deposit guarantee scheme possibilities out to the Minister of Finance on the Friday afternoon.  It was a mad rush, with some uneasy negotiated compromises (and everyone’s particular hobbyhorse concern got its own mention). I was probably too close to it to tell, and noted I wasn’t that comfortable with it, but when I got Alan Bollard’s signature he indicated he was happy with it.  I noted “lots of small details to sort out next week –  we hope only that, not implementation”.     To this point, we were focused mostly  on retail deposits, but I see in my diary that in The Australian on the Saturday there was talk from bank CEOs of a possible need for a wholesale guarantee scheme.

The full, unredacted, paper we wrote is available on The Treasury’s website.   The thrust of the advice was that (a) action was not necessary immediately, but (b) that should conditions worsen a scheme could be put in place at quite short notice.  The rest of the paper outlined the relevant issues, and the recommended features of any such scheme, and we advised against announcing a scheme until the remaining operational details had been sorted out, something we suggested could be done in the folllowing week.

These were the key features we suggested, largely accepted by the Minister.

dgs 1

One thing that puzzles me looking back now is why we were focused on guarantee options, rather than lender of last resort options.  The latter would have involved lending on acceptable collateral to institutions that we judged to be solvent, perhaps at a penal rate.  It was the classic response to the idea of a contagious run –  troubles elsewhere in the financial system spark concerns about other institutions, and people “run” –  cashing in deposits, retail or wholesale –  just in case.  A sound institution could, in principle, be brought down very quickly by such a run (empirically there are few such examples –  most actual runs end up being on institutions that prove to be at-best borderline solvent).

In the paper we sent to the Minister on 10 October we don’t seem to address that option at all.  I presume the reason we didn’t was twofold.  First, guarantees were beginning to proliferate globally.  And second, there probably is a pretty strong argument that if (a) you are convinced your banking system is sound, and (b) there are nonetheless doubts in the wider environment (in this case, a full scale global crisis, and a domestic recession), a guarantee is likely to be considerably more effective in underpinning confidence.  Not so much depositor confidence, as the confidence of bankers (and their boards).    Even if lender of last resort funding, on decent collateral, had been available without question, few bankers would have been happy to rely on that, and many would have been very keen to cut exposures, pull in loans, and reduce their dependence on the good nature of the Reserve Bank Governor.   A guarantee –  where the Crown’s money is at stake –  is a much stronger signal than a loan secured on the institution’s very best assets.   On the other hand, as the paper does note, once given a guarantee may not leave one with much leverage over the guaranteed institution.

Almost all of the subsequent controversy around the deposit guarantee scheme related in one form or another to one key choice.

All the systemically significant financial institutions in New Zealand were banks (not that all banks were systemically significant).  But they were not, by any means, the only deposit-taking institutions, and we were in the midst at the time of a finance company in which many companies were proving to be insolvent and failing.  Other finance companies appeared –  not just to the Reserve Bank, but to the market, and to ratings agencies – just fine.

Treasury and the Reserve Bank jointly recommended to the Minister that any deposit guarantee scheme include finance companies.  Why did we do that?

The simple reason was one of both fairness and efficiency.  Had we proposed to offer a guarantee only to banks (let alone only the big banks) then in a climate of uncertainty and heightened risk, there would have been an extremely high risk that such an action would have been a near-immediate death sentence for the other deposit-taking institutions, including ones with investment grade ratings, and in full compliance with their trust deeds.    We knew that finance companies (while small in aggregate) were riskier than our banks, but that was no good reason to recommend to the government a model that would have killed off apparently viable private businesses.  It still seeems, with the information we had at the time, an unimpeachable argument.  Classic lender of last resort models, for example, don’t differentiate by the size of the borrowing institution.

We weren’t naive about the risks –  including that there was still no prudential supervision of finance companies and the like –  and we explicitly recommended that risk-based fees (tied to ratings) be adopted, and the maximum coverage per depositor be much lower for unrated entities.   We included in the table an indicative fee scale, based credit default swap pricing for AA-rated banks in normal times, scaling up (quite dramatically) based on the much higher default probabilities of lower-rated entities.

We even included a indicative, totally back of the envelope, guess as to potential fiscal losses –  drawing on the experience of the US S&L crisis.  As it happens, actual losses were to be less than that number, even though the scheme as adopted by the Minister of Finance was less good than the one we recommended.  (Treasury provided some other –  but lower – loss estimates a few days after the actual announcement, but I can’t see those on the Treasury website and can’t now recall the approximate numbers.)

But all that was just warm up.   We’d been under the impression that the Prime Minister was going to announce, in her campaign launch speech, that preparatory work was underway on a deposit guarantee scheme.  That was probably her intention.  But that didn’t allow for the Rudd effect.  The Australian Prime Minister decided that he was going to announce an actual retail guarantee scheme for Australia that day –  the Sunday.  And so it was concluded that New Zealand had little choice but to follow suit.   As a matter of economics, there probably was little real choice but to follow the Australian lead.  But the timing was all about politics.  Neither economic nor financial stability would have been jeopardised if we hadn’t had a deposit guarantee scheme announced before the banks opened on Monday morning.  We’d have been much better to have taken a bit more time and hashed out some of the details with the Minister in his office in Wellington, not at campaign launches and then, as the day went on, airport lounges (at one point late that afternoon I –  who’d talked to the Minister perhaps twice in my life previously –  was deputed to ring Dr Cullen and get his approval or some detail or other of the scheme).   But I guess it might have left open a brief window in which critics might have suggested that New Zealand politicians were doing less for their citizens and their economy than their Australian counterparts.

The main, and important, area in which Dr Cullen departed from official advice was around the matter of fees.   We’d recommended that the risk-based fees would apply from the first dollar of covered deposits (as in any other sort of insurance).     The Minister’s approach was transparently political –  he was happy to charge fees to big Australian banks (who represented the lowest risks) but not to New Zealand institutions (including Kiwibank).  And so an arbitrary line was drawn that fees would be charged only on deposits in excess of $5 billion.   Apart from any other considerations, that gave up a lot of the potential revenue that would have partly offset expected losses.  The initial decision was insane, and a few days later we got him to agree to a regime where really lowly-rated (or unrated) institutions would have to pay a (too low) fee on any material increases in their deposits. A few days later again an attenuated pricing schedule was applied to deposit-growth in all covered entities.   But the seeds of the subsequent problems were sown in that initial set of decisions.

The weeks after the initial announcement were intense.  We rushed to get appropriate deed documents drawn up, dealt with endless request from institutional vehicles not covered who sought inclusion (property trust, money market funds etc), and set up a monitoring regime.  In parallel, we quickly realised that the way wholesale funding markets were freezing up suggested that a wholesale guarantee scheme was appropriate, and got something announced in a matter of weeks –  a much more tightly-designed, better priced scheme, operating only on new borrowing (but I’m biased as that scheme was mostly my baby).  As it happens, that scheme provided the leverage to actually get the big banks into the deposit guarantee scheme.  Once the government had announced the retail scheme the big banks had little incentive to get in –  they probably thought of themselves (no doubt rightly) as sound and as too big to fail –  and the scheme was an opt-in one (we couldn’t just by decree compel banks to pay large fees).   But the Minister of Finance –  probably reasonably enough –  insisted that if banks wanted a wholesale scheme (which they really did) it would be a condition that they first sign up to (and pay for) the retail scheme.  Perhaps less defensible was the Minister’s insistence that any bank signing up to the guarantee scheme indicate that it would avoid mortgagee sales of home owners in negative equity but still servicing their debt (the ability of banks to do so is a standard provision of mortgage documentation).

After the first few weeks of the retail scheme I had only relatively limited ongoing involvement, and so I’m not going to get into litigating or relitigating the South Canterbury Finance failure, and whether –  even the constraints the Minister put on –  and how that could by then have been avoided (the Auditor-General report some years ago looked at some of those issues).   The outcome was highly unfortunate, and expensive.  Nonetheless, it is worth remembering that the total cost of all the guarantee schemes – retail and wholesale – was considerably less than officials had warned was possible.  And it is simply not possible to know the counterfactual –  how things might have unfolded here had either no guarantees been offered, or if the finance companies and building societies had been excluded from day one.  Personally, I think neither would have provided politically tenable, but we’ll never know that, or how that alternative world played out.

But with the information we had at the time –  including, for example, the investment grade credit rating for SCF (which had outstanding wholesale debt issues abroad –  and actually my only meeting with SCF was about their interest, eventually not pursued, to try to use the wholesale guarantee scheme) –  the recommendation made on 10 October seem more or less right. Given the same information I’m not sure I’d advise something different now.  And once Australia had made the decision to guarantee retail deposits, there was little effective economic or political choice for New Zealand.   Had they not done so –  and there was real data, regarding increasing demand for physical cash in Australia, supporting Rudd’s action (rushed as timing was) – perhaps we could have got away with a well-designed wholesale guarantee only.   That would have been a first-best preferable world, but it wasn’t the set of facts we actually had to work with.