Dismantle and start again

As I previewed here, I recently had the honour of being parachuted into the crucible of the US policymaking debate when I was invited by the Rockefeller and MacArthur Foundations to present at the private Transforming America’s Housing Policy summit for Obama Administration officials in New York.

While in New York I was overcome by an unearthly feeling that the world I perceived was conspicuously different to that which my US colleagues could see. In trying to work out why, for instance, Australia’s financial system has to date been in such radically better shape than its US counterpart, I began to realise that there is a fundamental frailty that rests at the heart of the US financial architecture, which sets it apart from almost all other developed countries, and which has been largely responsible for both precipitating the current crisis and subsequently propagating it around the rest of our increasingly interconnected world.

In spite of all the (usually conservative) rhetoric about the problems with the government-sponsored enterprises (GSEs), Fannie Mae and Freddie Mac, I have heard no discussion of this much more far-reaching flaw sitting in the foundations of the US credit creation system.

The problem is a simple one: the vast bulk of all home loans in the US (around 70 per cent) are funded not using the balance-sheets of large transnational banks, and in turn the regionally diversified deposits of their customers, but via the far more complex, unstable and sometimes conflicted process of “securitisation”.

In contrast to the rest of the world – where securitisation, if it exists at all, has been a small yet valuable part of the housing finance mix – the process completely dominates home loan funding in the US. The concern is not with government support for homeownership, which one can find in justifiable forms in most countries, but the deleterious consequences of a financing system that initially evolved from the designs of competing states within the highly fragmented US federation, and which was distorted further by the federal government’s policy responses to the banking failures during the Great Depression. Of course, such failures were themselves an artefact of the exceptionally decentralised financing system that was the product of state-asserted control.

The outcome of these state and federal government decisions, and the continued missteps of US policymakers ever since (including the partial privatisation of Fannie Mae in 1968 to remove its debts from the government’s balance-sheet, and the misplaced creation of the second major GSE, Freddie Mac, in 1970 purely to compete with Fannie), has been the effective disintermediation of deposit-taking organisations as a source of housing finance in the US in favour of the GSEs and, crucially, the process of securitisation that they pioneered.

It is my belief that the two GSEs basically became a synthetic surrogate for the nationally-integrated banking systems that serve as the foundation for housing finance in most other countries. By doing so, they also attenuated pressure on state and federal governments to facilitate the wholesale consolidation of the geographically fractured and prone-to-fail US banking system that should have organically occurred over the 20th century.

When a mortgage is securitised, the lender that originates the loan does not keep it on their balance-sheet and hold it to maturity – the whole point of this exercise is to take assets “off balance sheet” and sell them to third-party investors, thus enabling the lender to recycle the original capital and embark on a new round of financing. Managed with the right regulations and controls for conflicts of interest, this makes considerable economic sense, as Joshua Gans and I have argued in the past.

Indeed, most regulators around the world have noted approvingly over the years that securitisation allows lenders to alleviate balance-sheet stresses and share some of their risks with third-parties, such as pension funds, who in turn get exposure to typically low volatility “mortgage-backed securities” that yield higher returns than conventional cash instruments. The very low risk and robust long-term performance of securitised home loans inAustralia has been testament to the merits of this diversification process for consumers, lenders and investors.

Trouble nevertheless arises when you create artificially strong incentives, as US governments repeatedly did, to predicate your entire housing finance edifice on securitised forms of funding to the detriment of the traditional deposit-taking market. Over and above stunting the growth of a nationally-integrated banking sector, synthetic incentives that instill securitisation as the preferred source of funding expose the overall financial system to a number of potentially destabilising conflicts. The most obvious of these is that the organisations that source new home loans and which are responsible for assessing their credit risk are removed from the institutions that ultimately own the assets and hence bear that risk. We have, therefore, a classic principal-agent problem.

Because the quasi-private GSEs had an artificial capital-raising advantage wherein investors imputed to them the US government’s AAA-rated credit rating (on the assumption they were backed by the state), they could source funds to underwrite home loans, and insure mortgage-backed securities, much more cheaply than their competitors (i.e. the non-AAA rated deposit-taking banks). The GSEs also had other crucial advantages such as tax exemptions and, most importantly, substantially lower capital requirements that allowed them to assume spectacularly higher levels of leverage than any normal bank. In fact, Fannie and Freddie developed investment banking-like leverage ratios of 20:1 and 70:1, respectively (these ratios rose radically if you included all the off balance-sheet mortgage-backed securities they guaranteed). Prior to the credit crisis materialising, the two GSEs alone funded or guaranteed around half of all US housing finance, with total on- and off-balance sheet liabilities of circa $5 trillion, which is terrifying when compared with the $9.5 trillion of US government debt at the time of their “conservatorship” (weasel words for nationalisation).

In the early 2000s, an additional wrinkle emerged when the GSEs were asked by, ironically, the Bush Administration to increase financing for low- to moderate-income regions with high minority populations which, in concert with shareholder calls to improve their returns, resulted in them using their AAA-ratings to invest for the first time in, or guarantee, higher risk “alt-A” loans. These loans had little borrower documentation, lower credit scores and/or higher loan-to-value ratios.

By 2008, the two GSEs held on balance-sheet or guaranteed around $1.6 trillion of these “non-prime” mortgages (or a remarkable 34 per cent of their total exposures), which unsurprisingly accounted for 90 per cent of their losses. Once again, the GSEs were effectively crowding-out non-AAA rated private lenders from their traditional alt-A domain and shunting them further down the credit curve. The not unpredictable consequence was a concurrent increase in even riskier (sub-prime) lending, which doubled from 10 to 20 per cent of all new US home loan origination between 2001 and 2005.

In many other countries, like Australia and New Zealand, where these quasi-government entities do not dominate housing finance, traditional banks ordinarily account for up to 80-90 per cent of all home loans with the vast majority of these assets retained on the lenders’ balance-sheets. And when lenders do securitise, they usually apply the same credit-assessment standards to the loans that are securitised that they use with the assets that are retained on their balance-sheets. Accordingly, in most western economies these lenders control the credit assessment process, they service the assets, and they bear the risk if borrowers default on their loans (and when they do securitise, they continue to service). In the US, all three activities – origination, servicing, and funding – are often separated, which in the absence of mitigating regulation leads to inevitable conflicts of interest. This is almost certainly one of the reasons why, for instance, Australian mortgage default rates are less than 15 per cent of US levels.

So how has the world’s largest and purportedly most advanced economy ended up with this unusual credit creation system? It seems that few people are seeking to address this fundamental question. I believe the explanation lies in two intertwined factors: firstly, the incredibly dispersed and granular US banking industry (there are currently over 8,400 banks and savings institutions, around half of which have less than $100 million in assets), which is disposed to frequent failure and even today has amazingly produced just one truly national coast-to-coast bank (Bank of America); and secondly, the US government’s panoply of interventions in response to the problem of persistent bank failures and the collapse of the financing system during the Great Depression.

The extraordinary degree of direct government involvement in the US housing and financing systems is hard for an outsider to fathom. In addition to the GSEs, the US Government created the largest public mortgage insurer in the world in 1934, known as the FHA, which specialises in insuring the losses of private lenders that source higher risk, non-prime home loans with deposits of less than 20 per cent and/or poor credit scores. Today, the FHA’s market share of new mortgages has risen from just 4 per cent in 2006 to nearly 20 per cent in 2008, with analysts predicting that it will hit 30 per cent in 2009.

Don’t misunderstand me here – as Joshua Gans and I forcefully argued in early 2008, governments do have a vital role to play intervening when critical markets fail, as they have done repeatedly during the credit crisis, to supply the “public goods” of a minimum level of liquidity and price discovery. But these government actions should not be allowed morph into institutions that permanently oppress private market activity, as was the case with the US’s deposit-taking system.

When I described the Australian housing market’s characteristics to the 200-300 policymakers at the summit you could see their jaws metaphorically hit the floor. Without any permanent government interventions of the kind seen in the US, Australia has generated a higher 70 per cent rate of home ownership, no sub-prime or alt-A market to speak of, no bank failures, no nationalisations, no real mortgage credit rationing, and current and long-term mortgage default rates that are a fraction of US levels.

The audience was even more amazed to learn that over 75 per cent of all Australian borrowers are on “adjustable rate” home loans, which are viewed as devil’s breath in the US (and those on fixed-rate loans only fix for one to five years). Significantly, these variable home loans’ interest rates are set based on the central bank’s target rate. In the US, however, the government’s post-depression actions set up a system where around three quarters of all borrowers have 30-year fixed-rate mortgages, which do not price off the central bank’s target rate but rather seven-year Treasuries, over which the central bank has limited to no control.

While a 30-year fixed rate mortgage sounds wonderful from an affordability perspective, as it did to policymakers during the depression, it has severely undermined the ability of the Federal Reserve to manage economic activity and the housing market in particular. Between August 2007 and December 2008, the Fed slashed its target policy rate by 500 basis points. Yet the average interest rate on outstanding US home loans fell by only 15 basis points. It has been no surprise, therefore, to see US default rates continue to rise in spite of the Fed’s efforts. In comparison, Australia’s central bank has been able to seamlessly deliver a 40 per cent drop in the cost of the variable rate home loans paid by most borrowers with 275 of its 300 basis points worth of rate cuts passed on by lenders to these households.

The clear point of difference between Australia and the US is government policy. While both nations are federations, the US was arguably disadvantaged by virtue of its far greater decentralisation across 50 states. All US banks were initially chartered and regulated by their individual states, which often had conflicting approaches. This made it near impossible for banks to pool risks and transfer liquidity from one state to another during times of crisis. The result was a high propensity for bank failures during the 19th and 20th centuries.

Following a bout of bank failures in 1907, Congress established the Federal Reserve in 1913 with a principal purpose of facilitating liquidity between banks to prevent these crises from occurring. Yet up until the early 1990s, laws remained in the US that prohibited banks from setting up branches in outside states and from merging with one another. As Charles Calomiris has observed, “economic logic often took a back seat to special interest politics and, occasionally, to populist passions.”

The central bank’s presence also proved incapable of stopping a continued proliferation of bank failures due to heavy localisation and a smorgasbord of often inadequate regulators including the Fed, the OCC, the OTS, the FDIC, the National Credit Union Administration, and separate state bodies. According to the FDIC, an amazing 2,698 US banks and savings institutions failed between 1984 and 2003. And it was only in 1994 that Congress finally agreed to repeal most of the prohibitions on interstate banking, yet by that time the damage had already been done. The incredibly fractured US savings system, subordinated as it was in the provision of housing finance to the unnaturally competitive GSEs, had been cast in stone.

The first-order cause of the global financial crisis was not the advent of sub-prime lending, greedy investment banks, non-recourse lending, community reinvestment acts, or the GSEs per se (although a case can be made that all of these things contributed). The underlying driver was over a century of flawed political decision-making that created a deeply dysfunctional and inherently fragile system of housing finance under which US bank balance-sheets were displaced.

Since the 1930s, the deposit-taking industry has been supplanted from the provision of traditional mortgage credit by the AAA-rated GSEs, which entrenched the complex and sometimes unstable process of securitisation as the dominant source of finance, in contrast to most other countries where balance-sheet funding proliferates. The destruction of the alignment of interests found where the originator and funder of homes loans are one and the same, combined with the absence of regulation to ameliorate these intrinsic risks, led to an inexorable deterioration in credit standards and the development of unusually risky (i.e. sub-prime) products. These higher-yielding instruments, which were increasingly originated by third-parties to households with a low probability of being able to service them, came to account for an unprecedented share of all new flows.

The government-created, yet purportedly private GSE duopoly, which acted as a surrogate for a nationally integrated deposit-taking system, stunted the need for the geographically dispersed and intrinsically fragile US banking industry to naturally consolidate and insulate itself from failure over the course of the 20th century. These structural flaws in the US credit creation system were exacerbated in the early 2000s, when the highly-leveraged GSEs, on the conflicted imprimatur of both government and shareholders, entered the much riskier nonprime segments of the US mortgage market, which ended up accounting for most of their losses. Once again, the traditional private lending sector was pushed further down the credit curve with a consequent explosion in sub-prime) loans.

As default rates inevitably rose and the system of securitisation instantaneously transmitted these risks to investors around the world, the dark side of capital market integration and globalisation emerged: liquidity in credit markets for completely unrelated assets collapsed as information asymmetries and irrational herd-behaviour propagated an extreme and often difficult-to-justify rise in global risk aversion. Defective mark-to-market accounting standards, premised as they were on the belief that “efficient markets” always priced assets accurately (but rarely during times of crisis), entrenched a vicious negative feedback loop as artificial declines in collateral values forced banks and investors all around the world to stop lending, triggering further reductions in asset values, and yet another contraction in lending, ad infinitum.

One overlooked point here is that many of the so-called “toxic” assets that commentators wax lyrical about were not toxic at all: the market failures triggered by the implosion in the US housing finance system precipitated illiquidity for all forms of credit internationally, which only then embedded the deleveraging death-spiral that destroyed asset values and which has now claimed much of the international banking system.

Today, the private lending market in the US has all but disappeared with the GSEs and the FHA accounting for an astonishing 95 per cent of all housing finance. Indeed, what remnants of private banking that do remain are being gradually nationalised with the US government now the largest individual shareholder in Citigroup and Bank of America.

My advice to President Obama, Timothy Geithner, Shaun Donovan, and Austin Goolsbee is that merely applying myopic bandages to the symptoms of these problems, and reinvigorating the GSEs, is emphatically not the long-term answer. The entire system of housing finance needs to be transformed to a bank-based balance-sheet focus. At some point, the GSEs should be fully nationalised and, alongside other public housing finance agencies, phased out of the day-to-day private credit infrastructure. Governments have a role to play supplying the public goods of liquidity and price discovery when markets fail – but only when markets fail.

In the long run, the Obama administration needs to create something that has been beyond previous governments: a robust and nationally-integrated private banking infrastructure, which is funded mainly through retail deposits, in order to firmly reposition balance-sheets as the principal repository of housing finance in the US. While there are undeniable diversification benefits to be garnered from securitisation, there is no evidence that the third-party capital available through this process should supplant the savings system.

*Christopher Joye presented to the private Transforming America’s Housing Policy summit for Obama administration officials in New York. Joye is managing director of research group Rismark International.

http://www.businessspectator.com.au/bs.nsf/Article/Joye-$pd20090226-PM69X?OpenDocument&src=sph

4 Responses to "Dismantle and start again"

  1. NFrazier   March 12, 2009 at 7:19 pm

    Here’s another possibility: use the fiscal stimulus to start a new bank.* Such a plan might:1. Circumvent the legal, accounting, and solvency snafus ensnaring the current nationalization attempts.2. Quickly provide liquidity to good businesses while letting bad ones fail in short order. (Obviously, only the most stodgy credit analysis and loan approval standards should be rigidly adhered to here. Additionally, an independent regulatory agency should be created to prevent the moral hazard of funneling taxpayer capital to corporations with aggressive lobbies…)3. End moral hazard in the US banking system and reimburse the US tax payer. In other words, if a bank relaxes lending standards with the expectation of getting bailed out later, the tax-payer instead creates a new entity to compete with that institution precisely at its moment of greatest weakness. As market share gets transferred over to this government bank, the other bank is forced to sell its lending facilities at distressed prices to the government bank. This new government bank could be privatized at a hefty profit to the tax payer at a later time.4. Help prevent the US economy from falling into a fiscal trap, prevent a run US treasuries, keep interest rates low, prevent the property crisis from worsening, etc.*Thanks to Mr. Rogers for this suggestion.

  2. Columbia academic   March 13, 2009 at 1:47 am

    This is the best and most original analysis of the origins of the GFC that I have ever seen. Well done and congrats to RGE for bringing this to my attention. Brilliant.

  3. Stanfordecon   March 13, 2009 at 4:10 am

    Best insights of 2009, I would say. It is a pity more American policymakers are not listening. Christopher, congratulations.

  4. Frank Mahlab MBA   March 18, 2009 at 6:01 pm

    This excellent article by C Joyce is one of the few addressing fundamental systemic faults. Writing from Melbourne Australia I can confirm his description of our systems. Whenever a lender is protected from risk in any financing a danger exists. In Australia virtually all home and most commercial loans are on a recourse basis. The USA system allows borrowers to walk away. In addition the complex insurance systems( eg AIG)failed to correctly assess the risk.C.Joyce did not address an alternate system to finance the poor into home ownership which has enormous stabilizing social benefits. For example Australia has first home grants from the government.