How much equity do banks voluntarily choose?

The Reserve Bank Governor is proposing that banks should have to fund a much larger share of their balance sheet with equity rather than debt.  So weak is our system that one unelected person –  on this occasion, one without much specialist expertise – can, more less on a whim, compel bank shareholders to either put billions of dollars more into their businesses, or to markedly downsize those businesses (most probably some combination of the two).    So long as the Governor jumps through the right procedural hoops, there are no appeals, and there is little formal public accountability.     Perhaps political pressure will mount on the Bank?  In these circumstances, one can only hope so.

Today I wanted to focus on a couple of strands in the arguments often used by academics and regulators to support the case for higher minimum capital requirements.   (Upfront I should say that they were arguments I used to be persuaded by, and to deploy myself.)  The post is a bit discursive: I’ve partly used writing it to help clarify my own thoughts.

Champions of higher bank capital ratios often note that bank capital ratios a hundred (or more) years ago were materially higher than those we observe today.    A good example of what they have in mind is this chart, taken from a speech a few years ago by a senior Bank of England official.

capital ratios

These aren’t, typically, risk-weighted capital ratios, just capital as a percentage of total assets.

One argument tends to run along the lines that if bank capital ratios were so much higher then, it could hardly be harmful for them to be much higher now.  And that claim is buttressed by the observation that deposit insurance (and a more general tendency for governments to bail-out all creditors of large banks) should have tended to reduce the  capital ratios banks would choose to run, in ways that are not necessarily socially desirable.   Thus, it is argued, one can’t deduce anything from observed capital ratios in recent decades since they, so it is argued, just result from subsidies (implicit or otherwise) to banking.   These sorts of arguments are made extensively by, for example, Professors Anat Admati and Martin Hellwig in their 2013 book The Bankers’ New Clothes  (Admati and Hellwig have both been recent professorial fellows at the Reserve Bank, Admati in 2016 and Hellwig late last year).

In principle, there is clearly something to the argument.   Were governments to guarantee all bank creditors full and timely payment of all their claims, and impose no obligations on banks (shareholders or management) in return, some banks might well rationally (if dishonourably) choose to run with hardly any capital at all: all losses would be borne by the taxpayer and all gains would be captured by the shareholders and promoters.

That, of course, isn’t the world we live in.  Even in places where there is deposit insurance (these days, most advanced countries other than New Zealand), wholesale and commercial creditors aren’t guaranteed.   And even when there is a strong too-big-to-fail presumption around some institutions (probably including the big banks in New Zealand) that presumption doesn’t apply to all institutions –  and even when the presumption exists, it is never held with certainty.

Many writers in this area come from countries with pretty comprehensive deposit insurance (from a depositor perspective, the system is the same whether you bank with Wells Fargo or some one-branch bank in the middle of nowhere).  But that isn’t so in New Zealand, so we should still be able to garner some useful information from the choices made around banks where there is, almost certainly, no strong expectation of a bailout.   This chart is from the Reserve Bank’s dashboard.

capital ratios dashboard

The regulator-required minimum capital (to risk-weighted assets) ratios are all the same (with the temporary exception of Westpac, after the never-adequately accounted for episode discussed here).   But what is striking is that there is no systematic difference between the actual capital ratios of the big-4 locally incorporated banks and those of the four smallish New Zealand owned banks.

Perhaps you see things differently, but I’d assume that depositors (and probably all other creditors) in the big-4 would end up fully-protected in any event much short of the physical destruction of New Zealand.  Between parent support and the political imperatives, that seems a pretty safe bet (at least a 95 per cent chance).  Neither the parent nor the government is charging upfront for that likely support.  I’d also assume Kiwibank creditors would be bailed out, for a slightly different mix of reasons (it is still 100 per cent government-owned).   And, on the other hand, the chances of the Crown bailing out creditors of the four small New Zealand banks  –  other perhaps than in a crisis that was taking down the whole system –  seem much much smaller.  It might even be credible to suppose that the OBR tool would be used if one of those banks were to fail.

And yet there is a striking similarity in the capital ratios across all these banks (and a striking similarity in the margins above minimum regulatory requirements). It doesn’t appear that consistent with a story in which the big banks are now able to get away with artificially low levels of capital because of the actual or implied bailout and guarantee risks.   And that is especially so when one recognises that none of the four small local banks has a deep-pocketed parent who might be prevailed on to recapitalise the bank if it got into trouble.

There is a caveat to this comparison.   The smaller banks (including Kiwibank) have to use the standardised approach to calculating regulatory requirements, while the big four are allowed to use (a constrained form of) their own internal ratings-based models.    For many assets, the latter approach will result in a bank being required to use less equity funding (for the same actual loan) than is required under the standardised approach (narrowing that gap is one aspect of the Reserve Bank’s current consultation).  If the big-4 banks were required to report on the same basis as the other banks, their reported capital ratios would be somewhat lower, but it wouldn’t change the fact that none of the smaller New Zealand banks have actual total capital ratios now as high as what the proposed Reserve Bank requirements will involve in the future (taking account of the buffers above regulatory minima banks will choose to maintain).

So how do we think about the long-term historical experience, of the sort captured in that earlier Bank of England chart?  If we go back 140 years, British banks in this sample had ratios of equity to total assets of around 15 per cent, and US banks something more like 23-24 per cent.

If we are thinking about a small listed New Zealand bank mostly doing business lending, I was interested to see from Heartland Bank’s latest disclosure statement that total balance sheet equity was about 14.7 per cent of total balance sheet assets.

But if we are comparing big New Zealand or Australian banks to those in the US or the UK in 1880, we are really comparing apples and oranges.   The biggest single difference is the predominant nature of the credits.   The biggest single chunk of loans on the balance sheets of New Zealand banks now are residential mortgages (followed by farm mortgages) –  secured by the considerable collateral of the underlying assets. Nineteenth century banks were typically didn’t lend to house purchasers (or farm purchasers) –  for that matter, 1960s trading banks in New Zealand didn’t either.      Loan losses on diversified portfolios of residential mortgages loans are not typically high.  My supposition –  I haven’t checked this story out –  is that 19th century banks will also have typically had larger peak exposures to a small number of borrowers.

In the US in particular, geographic diversification was often almost impossible to achieve (branch banking restrictions and all that).  You might reasonably respond that New Zealand is small, but even today New Zealand has slightly more people than the median US state, and most US banks were historically more tightly constrained than even a single state (one of the reasons they’ve had repeated banking crises and, say, Canada hasn’t).

And one could add that 19th century Britain and the United States were countries with fixed exchange rates, the US wasn’t long out of a civil war, and other aspects of its monetary system made its banks prone to liquidity crises.  The experience of the last 50 years or so –  with floating exchange rates –  is that countries with fixed exchange rate have more difficulty coping with economic shocks than countries with floating exchange rates.    We’ve seen that most recently in places like Ireland and Spain.

It may also be relevant to note that 140 years ago concepts of limited liability (including in banking) were still relatively new.  It takes time for the market (broadly defined) to work out what financing structures are sustainable and appropriate, balancing risk and opportunity.

We have another class of financial intermediaries in New Zealand that, almost certainly, no one expects would be bailed out if they got into trouble.  Non-bank deposit-takers  are now under the regulatory oversight of the Reserve Bank, with minimum capital requirements imposed by the Minister of Finance by regulation.  But these are minima only.  If there is no credible expectation of a bailout if things go wrong, any capital ratio materially above the regulatory floor must presumably reflect the judgement of shareholders, depositors, managers etc about the best (for that firm) mix of debt and equity.  I don’t have time or energy to go through the accounts of all the NBDTs, but I dug out those for the Nelson Building Society, a longstanding entity that mainly does residential mortgage lending, and without a huge amount of geographic diversification.  They report a risk-weighted (using RB approved weights) capital ratio of 10.09 per cent (equity is 6.7 per cent of unweighted total assets).  NBS’s credit rating isn’t particularly high –  were I a depositor I’d probably be keen on them having a higher capital ratio – but for a small not-overly-well diversifed lender, there is no sign of the market demanding anything like the 20 per cent (risk-weighed) capital ratios that will be implied by the Reserve Bank’s current proposals.  (Checking another building society, the Wairarapa Building Society reports an 11.7 per cent risk-weighted capital ratio.)

Finally, I suspect a great deal of the push for higher capital ratios is at least buttressed by the Modigliani-Miller story.   One of the Bank of England papers in this area (one of those cited by the Reserve Bank) gives this nice summary of the idea

Modigliani and Miller (1958) showed that, under certain assumptions, moving to higher levels of funding in the form of common stock, and therefore lower levels of debt and financial leverage, would leave the total cost of funding unchanged. In particular, the Modigliani-Miller (MM) theorem implies that as more equity capital is used, return on equity becomes less volatile and debt becomes safer, lowering the required rate of return on both sources of funds. It does so in such a way that the overall weighted average cost of funds remains unchanged. This idealised situation represents the case where there is a complete (100%) offset in relative funding costs as the debt and equity compositions change.

In other words, financing structures don’t really matter much (on certain assumptions), leading to a view that it doesn’t really matter much then if officials and regulators insist on one financing structure (a large share of equity) over another.   Pretty much everyone accepts that Modigliani-Miller (MM) doesn’t hold perfectly, but equally that it does hold to some extent (all else equal, over time, expected returns will be lower  –  and less volatile –  in a business with a greater share of equity funding).    One of the reasons MM often doesn’t hold fully in practice is the tax system: most tax systems tax equity returns more heavily than debt returns (often double-taxing equity returns, both when earned in the company and then when distributed to the owners).

But that wrinkle isn’t an issue for locally-owned institutions in either New Zealand or Australia (both countries run dividend imputation systems).  In other words, even if there is a systematic bias away from equity in other countries (including in the financial systems), there is no reason to expect to see it here, for locally-owned entities.  So when we look –  see above –  at the risk-weighted capital ratios (or simple ratios of capital to assets) for small New Zealand owned financial entities, not only is there little or no bailout risks factored into the chosen ratios, but there should be no tax system bias either.

More generally, if MM were the key insight on financing structures –  as distinct from being one element in a complex mix –  shouldn’t we expect to see capital ratios scattered almost randomly between (something close to) zero and a hundred per cent?  After all, the financing structure doesn’t affect the total cost of finance.  But that isn’t what we see at all.  Sure, there are some companies (even listed ones) that choose to have no (net) debt at all, although even that choice seems often to relate to anomalies in the tax system.  Financing structures tend to bunch by industry, suggesting in each case something about the nature of the industry itself influences those choices (something that won’t always be apparent to keen regulators).  That appears to be true for the financial sector as well –  even in parts of it where there is no credible bailout risk.

I’m not opposed to regulatory minimum capital requirements.  If governments are going to either provide deposit insurance, and/or behave in ways that create a perception of probable bailouts, some regulatory minima are almost certainly needed.   But where there is no deposit insurance, and there is little or no bailout risk, private market choices about financial structures in banking look as though they should tell us something about the economics of the industry.   All else equal, there might be a good case for ensuring that minimum ratios for banks that might be expected to be bailed out are in the ballpark of those intermediaries with no such (probable) assurances.  But in a New Zealand context, there doesn’t seem to be anything in the practice of those smaller institutions that would point in the direction of regulatory minimum capital requirements anywhere near as high as what the Reserve Bank is proposing.

Sadly, it probably won’t surprise you now to know that in their consultative document the Reserve Bank does not engage with these sorts of perspectives or experiences at all.

On another matter, I noted the other day that someone had started a Twitter account linking to various posts from this blog.   Having listened to the arguments of the (anonymous to me) person behind that account, I have activated a Twitter account of my own.  Not many of my arguments usefully reduce to 280 characters, so for now anyway I expect to use it mainly for links to new posts, or perhaps the occasional article that I think readers might find interesting or an old post relevant to some topic that has come to the fore again.