The author has been described by News Ltd as an "iconoclast", "Svengali", a pollie's "economist muse", and "pungently accurate". Fairfax says he is a "Renaissance man" and "one of Australia’s most respected analysts." Stephen Koukoulas concludes that he is "85% right", and "would make a great Opposition leader." Terry McCrann claims the author thinks "‘nuance’ is a trendy village in the south of France", but can be "scintillating" when he thinks "clearly". The ACTU reckons he’s "an enigma wrapped in a Bloomberg terminal, wrapped in some apparently well-honed abs."

Tuesday, February 16, 2010

The sharemarket sham

There is an investment category out there that you likely have a large chunk of your wealth tied up in. The problem is that it’s literally 11.6 times riskier than ‘cash’ (ie, putting your money in the bank). In fact, this asset-class has abruptly halved in value three times in the last 20 or so years. While it is usually the second largest investment in any super fund portfolio, it is almost perfectly correlated with the average fund’s other biggest allocation. That is, it offers little-to-no diversification (indeed, it merely parlays up your risks). On a risk-adjusted basis, it has been the single worst performing investment that we can measure. Some of you may be intuiting where I am heading: ‘global equities’, which combined with Australian shares accounts for more than one-half of the $1.2 trillion in Australian super savings. The scandal here is that there is scant analysis that rationalises superannuants assuming such enormous exposures to the world’s two riskiest asset-classes.

We recently updated our historical investment data (see table below), and the performance of global equities over the last 27 years has been nothing short of abysmal. And that’s on a ‘raw return’ basis—ie, with no accounting for ‘risk’. Employing a risk-adjusted metric (refer to the third row, where higher numbers denote superior outcomes), Australian equities also don’t stack up relative to fixed income investments, such as bank bills and government bonds. I am pretty sure one could also add AAA-rated Australian home loans and A1+ corporate debt to the fixed-income outperformers, although it is difficult to quantify their long-term returns due to a lack of suitable time-series data.

What is also interesting is that ‘net’ residential real estate (we’ve simply halved the gross yields) and long-term government bonds have bettered global equities and listed commercial, industrial and retail real estate (aka ‘LPTs’) in raw return terms. And although Australian shares have offered the highest absolute payoffs, they have done so with almost unimaginably high risk: the annual volatility of Australian equities has been an extraordinary 19.7 per cent. That’s 2.3 times the riskiness of government bonds, 5.6 times the volatility of residential property, and 15.5 times the risk of cash.

Unfortunately, most of us underestimate risk (including many supposedly sophisticated investors), and focus obsessively on returns. It is no different to the psychological biases that push us to throw away money at the racetrack or casinos, notwithstanding that the expected outcome of these endeavours is negative. And that touches on a sobering fact that one should never lose sight of: risk represents the probability of loss. This is precisely why any person who tells you that shares are ‘the best place to be’ is mad: the only way they can possibly arrive at this conclusion is by completely ignoring risk, or by assuming that you are trying to generate unusually high returns. Most of these folks are no better than Macau junket operators.

The remarkably high risk of equities was practically borne out in the circa 50 per cent peak-to-trough losses registered by both the Australian and global sharemarkets during the GFC. And this brings us to another oddity: for all the talk of ‘house price bubbles’, by far the biggest bubbles were actually in listed equities. More specifically, while the Australian, UK and US sharemarkets all fell in value by around 40-50 per cent, their house price indices declined by only 4 per cent, 14 per cent, and 30 per cent, respectively.

So why don’t we hear more about the great sharemarket bubbles of 2007-09? Were not almost all of the corporate collapses in Australia tied directly to equities (eg, Allco, Babcock & Brown, MFS, etc)? Was not the single biggest threat to our financial system posed by corporate debts, which dominated the (now quickly evaporating) provisions set aside by the major banks in anticipation of severe lending losses during the unfulfilled recession? Just like corporate, and notably not household, lending brought the Australian banking system to its knees in 1991. Have not most of the other major global calamities we’ve faced during the last three decades (ie, the ‘junk bond’ boom and bust, the associated 1987 crash, the LTCM bailout, and the 2001 tech wreck) all originated from, or, were transmitted by, the equities, corporate debt and/or commercial real estate markets? The poorly publicised truth is that we can answer all these questions in the affirmative.

Perhaps the collective equity market amnesia has something to do with the fact that while there are trillions of dollars of institutional capital tied up in shares, there is little direct institutional investment in housing. Just millions of faceless mums and dads.

One would have little appreciation for the risks inherent in equities reading the media, listening to financial planners, or by inspecting the asset-allocations of retail super funds. Super funds in particular have enormous and quite unjustifiable weights to Australian and global equities of around 50-60 per cent. Here one shudders to think what the performance of the equities market would have looked like had nearly one half of the $1.2 trillion worth of super savings not been ploughed into it following the establishment of compulsory super in 1992.

What makes the 50-60 per cent asset-allocation to Australian and international shares so worrying is that both these categories are highly correlated with one another. If you look at the table below you can see the number “0.64” sits at their intersection. The only higher number captures the correlation between LPTs and local equities, which is an unsurprisingly strong 0.68.

This basically tells us that over the last circa 30 years there has been a 60-70 per cent commonality in the movement between these sectors. Put differently, investing your money across Australian equities, global equities and LPTs offers almost no diversification. The real story underlying these numbers is actually much more damning since during severe market corrections (when you most want to rely on whatever diversification exists) the correlations converge to close to one. That is, the asset-classes move in near perfect unison, as we saw in 1987 and during the more recent GFC.

Threading all this together, if you are an investor targeting even a relatively high long-term nominal return, which naturally favours equities investments, of, say, 9-12 per cent per annum, and you have the choice of apportioning your money across all the abovementioned categories, a standard portfolio optimisation procedure teaches you two key things: first, most of your money should be committed to safer debt securities; and, secondly, whatever minority share of your portfolio that you do allocate to equities should likely be invested in one sector, rather than arbitrarily across all three as is common today.

It is worthwhile noting here that this 27 year time horizon is the longest period over which one can objectively measure all the asset-class returns. Now punters might allege that it has just been a bad epoch for equities. Yet the truth is that the ASX was only established in 1987, and longer-term equities data is subject to severe ‘survivorship biases’ whereby poorly performing and delisted companies are removed from the historical index datal. This only leaves ‘winners’, so-to-speak, and induces an upward bias in recorded equities performance.

These learnings are also important when one considers the recent statements of the Governor of the RBA. Glenn Stevens’ incantations on asset prices last week were widely misinterpreted by a typically unthinking commentariat as being only relevant to housing. While residential property is an unambiguously material risk in the system-stability matrix, the unavoidable empirical fact is that it ordinarily poses less of a concern for central banks than the corporate equities and debt markets. These are by far the more frequent delivery devices for catastrophic contagion.

Over and above the searing volatility of listed equities illustrated above, the hazards inherent in corporate debt and equities markets are evident in the capital treatment applied to them by regulators. If a bank invests in equities, it cops a full ‘capital deduction’: that is, it must have 100 cents of capital set aside for every dollar of equities exposure. If that bank lends to a business, it is subject to a 100 per cent ‘risk-weighting’, which means that they need around $8-$10 of capital for every $100 they lend out via corporate loans. Yet in part because residential property has dramatically lower risk (note the very low 3-4 per cent per annum volatility associated with a national portfolio of housing in the table above), and gives rise to only very small historical losses vis-à-vis business lending (as demonstrated by CBA’s long-term loss rates here), residential mortgage lending attracts a much lower 15-25 per cent risk-weighting from the regulators, which means the major banks only have to set aside $2-$3 of capital for every home loan they make.

Any journalists calling for central banks to ‘target’ house prices, which, by the way, Glenn Stevens has previously ruled out (the RBA is only really worried about coincident asset price and credit booms), is, by definition, also arguing in favour of using interest rates to manage equities. But for one reason or another we almost never read this line of reasoning.

When one evaluates more carefully the house price corrections in the US and UK between 2007 and 2009, it is easy to see that trying to influence them ex ante via interest rates would have been a very silly strategy. As the US Federal Reserve has quite convincingly shown, the housing bubble had little to do with the absolute level of mortgage rates in the US, and was largely attributable to the age-old boom-bust cycle in lending standards, which, in this most recent incarnation, was illustrated most strikingly with the development of the new market in ‘sub-prime’ lending.

The emergence of the US crisis also had a lot to do with their peculiar system of housing finance, which resulted in two quasi-public (and presumed-to-be-government-guaranteed), yet technically private, entities, Fannie Mae and Freddie Mac, supplying around half of all the finance for home loans. This government-established duopoly resulted in ‘securitisation’—as opposed to conventional ‘hold-to-maturity’ balance-sheet lending—being used to fund around 60 per cent of all mortgages in the US. In every other nation across the globe securitisation was but a small part of the financing mix, with the overwhelming majority of all lending being undertaken by banks that controlled the credit risk assessment process, and acted as the ultimate owners of the asset (securitisation divorces these two activities, which while yielding profound portfolio diversification benefits, also elicits non-trivial conflicts of interest, which were particularly pronounced in the US).

Had the Federal Reserve assaulted house price growth in the mid 2000s with higher interest rates, as some lazily argue they should have done, we would likely be living in a world today where the GFC was ascribed solely to its actions. Higher interest rates would almost certainly have triggered the spike in default rates that came to pass in any event. US house prices would have experienced serious falls, which would have likely brought about a similar collapse in the structurally-flawed SIV and CDO markets, and still propagated the sudden rise in global risk-aversion that asphyxiated liquidity in global funding markets. And, ironically, the Fed would still have been burdened with much of the blame for the ensuing crisis, albeit this time for causing it rather than failing to prevent it.

Today we would be listening to central bankers around the world wearily conclude that the Fed should never have departed from its first-order inflation focus, and that employing the very blunt interest rate instrument to vainly manage asset prices, which are merely manifestations of deeper dysfunctions, is, on the balance of probabilities, a recipe for disaster.

Rather than trying to crudely crush the symptoms of these problems with a tool that inevitably wreaks collateral damage across the wider economy—and, as Dr David Gruen at Treasury and Dr Guy Debelle at the RBA have wisely argued, would have a highly questionable ability to achieve these aims in any case—surely the smart strategy is to address the source of the malaise: that is, lending standards, which centuries of financial market history have shown to be predictably counter-cyclical.

If we could have our time again, would not the intelligent regulator try to directly combat the fundamental conflicts of interest in the US between the originators of home loans and the end holders of those assets? Only by attempting to cauterise the moral hazard spawned by this disconnect, and more generally improve the otherwise weak regulation of consumer finance in the US, would policymakers have had a genuine hope of minimising the high default rates associated with non-prime loans, which were the key driver of the distressed sales that were the signal characteristic of the crisis.

Higher interest rates would have been the worst possible remedy to the ticking time bomb that was the ballooning sub-prime bubble. Sure they might have helped prevent at the margin further credit growth and asset price inflation, and perhaps limited the extent of the ensuing deflation. But they would also have accelerated the rise in default rates and distressed sales that were the unavoidable result of poorly structured consumer finance products.

And which US agency bore part responsibility for regulating consumer lending? Why the Federal Reserve. So is not one of the more enduring lessons from the GFC that giving independent (ie, unaccountable) bodies like central banks formal regulatory powers beyond their inflation mandate a sub-optimal solution? This is exactly why we hived off APRA from the RBA—because the regulation of banks, insurance companies, and super funds needed to be undertaken by a dedicated agency that was directly accountable to the Treasury.

Central banks have been deliberately constructed as independent institutions only for the purposes of meeting their difficult price stability objectives, which are otherwise ripe for political interference. But this is a world apart from having regulators removed from the publicly-elected polity. The notion of an independent central bank possessing regulatory power is inconsistent with our parliamentary democracy, while making coordination with other regulatory agencies that are accountable to government intrinsically difficult.

Monetary policy would have also done next to nothing to prevent the UK crisis. The UK housing market correction, which was actually only mild with a circa 14 per cent decline in values, was brought about by the collapse of the global funding markets upon which banks like Northern Rock relied. This is why the fall in UK house prices preceded the subsequent rise in UK default rates; ie, it was the exact opposite of what happened in the US.

Almost all economists agree that the correction in UK home values was triggered by a credit crisis: lenders like Northern Rock could no longer fund themselves. So they stopped providing finance. And if credit markets collapse, asset prices are going to fall no matter how strong the underlying fundamentals may be (several UK government reviews had concluded that there was actually a significant underlying housing shortage).

This is also the main point of departure between Australia and the UK: for a whole bunch of largely fortuitous reasons, Australia’s banking system did not collapse, as it did in the UK, and did not, therefore, need to be nationalised. As a consequence, residential credit remained more or less freely available for high quality borrowers. So the logic that somehow the Bank of England could have prevented the UK crisis by using interest rates is offensively asinine. It overlooks the causes of the calamity, and diverts attention away from their longer-term remedies.

I agree with Glenn Stevens that monetary policy can help ‘condition’ the financing environment. And there is no doubt that prudent jaw-boning, or ‘open mouth operations’, can assist in drawing attention to the inherent cyclicality in credit markets, and the risk that the collapse in the availability of finance can have adverse real economy ramifications.

On the other hand, I think Glenn’s arguments are weak when he claims that macro prudential tools, such as pro-cyclical capital provisioning, that have the potential to directly address the underlying boom-bust cycle in credit markets, face the same challenges as monetary policy.

There is an essential difference between the two processes: whereas macro prudential instruments would be geared towards remedying the causes of credit crises, monetary policy can only very haphazardly affect the observable symptoms (ie, via influencing asset prices with long, variable and highly unpredictable lags).

Setting aside the counterfactuals I proposed regarding the US Federal Reserve above, another weak aspect in Stevens’ argument is that we have several empirical precedents that likely educate us as to what happens when central banks start trying to move beyond their inflation-targeting framework in order to focus explicitly on asset prices (and the formal conduct of monetary policy in Australia in 2003 is not one of them, no matter how hard folks might like to reinvent history). There is, for example, one strong school of thought amongst respected economists, such as Bernanke, Keynes and Friedman, that the primary catalyst for the Great Depression was the US Federal Reserve’s efforts to burst the sharemarket bubble.

As Stephen Kirchner has pointed out, in 2002 Ben Bernanke concluded that “the correct interpretation of the 1920s is not the popular one – that the stock market got overvalued, crashed and caused a Great Depression. The true story is that monetary policy tried overzealously to stop the rise in stock prices. But the main effect of the tight monetary policy…was to slow the economy. The slowing economy, together with rising interest rates, was in turn a major factor in precipitating the stock market crash.” Kirchner goes on to observe:

“The singular cause of the Great Depression of the 1930s, in Bernanke’s view, was that the Federal Reserve fell under ‘the control of a coterie of bubble poppers.’ Bernanke was merely reaffirming a well established consensus among economists, ranging all the way from John Maynard Keynes to Milton Friedman. In his A Treatise on Money, Keynes said ‘I attribute the slump of 1930 primarily to the deterrent effects on investment of the long period of dear money which preceded the stock market collapse and only secondarily to the collapse itself.’ Friedman’s 1963 A Monetary History of the United States also laid blame for the Great Depression squarely at the feet of the Fed and its attempt to become ‘an arbiter of security speculation or values.’”

Guy Debelle has also argued that Australia’s experience during the late 1980s reveals the type of monetary policy settings that would be required to bring asset and credit markets to heel:

“I do not think that a slightly tighter setting of interest rates would have prevented the development of the imbalances that have led to the current financial crisis. When human psychology is such that optimism about asset price rises is at the fore, then an excessively stringent setting of interest rates would be required to suppress the optimism. The Australian and Scandinavian experience in the late 1980s shows the sort of interest rate settings required to achieve such an outcome. In that example, a credit boom and bubble-like asset price dynamics took hold and only a very high setting of real interest rates ultimately curtailed that, but at the cost of a historically high level of unemployment. I do not think it would be socially acceptable or desirable to endure the level of unemployment that would come with the high interest rates necessary to pop the bubble. It is asking too much of the single monetary policy instrument, namely, the targeted short-term interest rate to target both financial excesses and inflation.”

My guess is that the first central bank that genuinely attempts to manage asset prices will also be the first central bank to lose its independence. And this is to say nothing of the far-reaching damage a failed attempt to influence asset prices would have on both the central bank's credibility, and the so-called 'nominal anchor' (ie, inflationary expectations). As one Fairfax journalist noted last week, such action in Australia would almost certainly require a rewriting of the existing agreement between the government and the RBA, which was only struck for the first time in 1996. Good luck trying.