Banks given breathing space but must still adopt FSA’s property risk model, writes Alex Catalano
Two years ago, the Financial Services Authority warned UK banks that it did not think their systems for calculating how much capital to put aside to cover real estate loans were good enough. Since then, lenders and the FSA have been tussling over slotting, the alternative system for calculating real estate lending risk that the FSA is now imposing on banks.
Real Estate Capital has discussed slotting with some of the people responsible for real estate lending in major UK banks, but they did not want to be identified. Slotting is a highly sensitive issue, not least because it has the potential to increase dramatically the amount of capital a bank has to hold to cover losses on its real estate loans.
The most alarmist estimates suggest that UK lenders would need to find another £20bn-£40bn of regulatory capital. The British Bankers Association and Property Industry Alliance argue that slotting is a crude instrument, which if applied now, will suck even more lending out of the market and increase pricing. At worst, it could trigger a flood of foreclosures and sales that would undermine property values.
The FSA is sticking to its guns and insisting that UK banks must use slotting. “It is the only thing available in the current regulatory framework, so that’s what we have to do,” says one banker.
However, the FSA has accepted that imposing a deadline for all banks to adopt slotting in its current form could destabilise the market. “There was a penny-dropping moment when the Bank of England saw this would be a very bad thing for the overall real estate market,” is the way one attendee described a meeting between regulators and industry representatives.
The regulators now acknowledge that there are problems with the slotting regime, after withdrawing revised guidelines issued in June 2011. Since January, the FSA has been reviewing banks’ risk weightings. It has also held talks with some UK insurers that are lending on commercial real estate, to ensure that it was covering all financial institutions in the market.
Banks buy some breathing space
Real Estate Capital has learned that the FSA has been working on a paper proposing a way forward, based on their review. “Now we hear the FSA is going to allow lenders to glide into this structure,” says one lender. “It is saying: ‘You must make the change, but you have two to three years to allow this to happen’. That gives us time to work out a lot of problem loans and issues.” Another lender notes: “The transition has begun. Banks are absorbing the cost of additional capital. It’s relevant to everything that’s happening.”
The clearest signal of this change comes from Nationwide, which Real Estate Capital understands is the first UK lender to apply slotting. Sources say the impact for the building society’s regulatory capital was quite substantial and contributed to its decision, announced in June, to pull out of big-ticket real estate lending and run its commercial real estate book down by £4bn-£5bn.
The amount of additional regulatory capital slotting may require of lenders will depend on how conservative the risk weighting of its loans was previously. Real Estate Capital has heard figures such as averages of 40% bandied about. “The degree of impact depends on where each bank starts,” says one banker. “But the aggregate effect is to require more capital behind the real estate book.
“The question for every bank is: ‘Do we consider that this is the appropriate amount of capital to have in a single sector?’ Slotting will play into the quantum of lending available in the market. That is my biggest concern.”
The problem for the banks is that they don’t have a better system to offer. Their internal models have been developed by specialists in quantifying risk, without much – or perhaps any – input from their real estate specialists. As one real estate lender puts it: “They’re quants; they don’t under-stand anything about real estate.”
A second and very real problem is that banks’ data for real estate loans is seriously flawed. Many do not have all the necessary information; in other cases, it is there, but IT systems cannot extract it in the right format. Or, the run of data may not be long enough – a robust model would incorporate several property cycles. The Bank of England asked the major UK banks to assess the impact of slotting on their loanbooks, but they “seemed incapable of doing this”, sources told Real Estate Capital.
There is no doubt that the banks’ internal ratings models were unfit for purpose. While recognising this, real estate bankers question whether anyone could have foreseen the scale of downturn. But is slotting any better? Until now, there has been no formal, quantitative study of the issue.
IPD quantifies the risks
However, IPD recently undertook research into whether risk weights calibrated by slotting were sufficient, insufficient or overly conservative to absorb losses during the last cycle. The compelling feature of this study is that it is completely independent of UK lenders and based on a very robust and granular set of data: billions of pounds worth of UK real state monitored by IPD.
Although these are institutional holdings, the data set includes the good, the bad and the ugly: so-called prime, secondary and tertiary assets, as well as all lengths of unexpired leases.
IPD ran simulations ‘originating’ hypothetical loans on these IPD properties, based on their actual values and net incomes at the peak of the market in 2007. The terms were assumed to be for seven years, needing refinancing in Q2 2014, with a fixed interest rate of 6% for the period.
IPD used only measurable criteria in allocating these hypothetical loans to slots: loan-to-value ratios (LTV), interest cover ratios (ICR), debt service cover ratios (DSCR), weighted unexpired lease terms, and the quality of the asset and location (see table). The latter was defined as an estimated rental value per square metre, relative to similar assets in similar locations.
The criteria IPD used are based on the revised guidance – now withdrawn – that the FSA issued in June 2011. The exercise tracked the defaults, as measured by LTV covenant breaches and shortfalls in DSCR/ICR over quarterly periods, to estimate the pattern of default that would have taken place during Q2 2007- Q2 2009 downturn. One scenario assumed bullet loans, a second, part-amortising ones.
It found that: the capital slotting required for strong and good loans was well above the estimated loss given default (LGD), even at the nadir of the cycle; the risk weights attached to each slot do not reflect the losses behaviour of the loans in IPD’s simulations; the criteria for debt service coverage ratios and interest coverage ratios are in conflict, so that amortising loans are treated more strictly than bullet loans, which works against conservative underwriting; the FSA’s draft guidance on lease length is too restrictive and out of step with the current market norm.
IPD concluded that slotting as currently structured is a flawed method of risk management. It might, IPD indicates, discourage real estate lending altogether and at very least will discourage low-risk lending.
However, a more refined system of slotting, with more slots and better calibrated risk weightings, does have the potential to guide lending in a stabilising way, it concludes. IPD plans a follow-on study that would explore this alternative.
Meanwhile, the lending industry is breathing a sigh of relief that regulators have backed off implementing their initial slotting revamp and are studying the issues. “We’ve averted a disaster in my eyes,” one banker says. But the problem now, is: “We’re in a bit of stasis at the moment. Only as we go further down the path with the FSA will we see what it looks like and where it comes out.”
Slotting timeline: the evolution of risk regulation
2007-08: Slotting is introduced in UK under the Basel II banking regulations
Slotting is a formula for calculating the regulatory capital banks must set aside for certain kinds of specialised lending: project finance, commodity finance, finance for ‘objects’ such as ships and satellites, and for income-producing real estate. It assigns risk weightings by classifying loans into one of five categories: strong, good, satisfactory, weak and in default.
This is done by looking at financial strength, including market conditions, loan- to-value ratios and cash-flow predictability. Asset characteristics such as a property’s location, design and condition are also taken into account, as is the strength of the sponsor or developer; not just financially, but their track record and relationships with others in the sector. Finally, the kind of security the bank has on the loan – assign-ment of rents, etc – is also considered.
Any sensible lender would consider these factors when making a loan. But crucially, slotting, as laid out in Basel II, did not attach any figures to the criteria it lays out, even for measurable ones, such as loan-to- value ratios. The FSA could have refined Basel II’s regime, but decided that it didn’t have the time and that it was unlikely to
“make a significant improvement on the work already done by Basel”.
However, Basel II also allows banks to develop their own systems – internal models for risk-weighting loans, providing these meet certain criteria for estimating the probability of default (PD) and the loss given default (LGD). Most of the main UK lenders have been using internal models on their commercial real estate loan portfolios.
October 2010: Financial Services Authority warns banks
The 2008 financial meltdown led regulators to focus on the amount of capital banks should hold to insure they can withstand market turbulence. The FSA Credit Risk Standing Group issues a ‘pre-consultation’ paper saying banks’ internal models for weighting real estate loans fall short of what is required to meet the minimum standards prescribed by Basel II. It warns that if they can’t come up with a compliant model, they will have to use slotting.
June 2011: FSA consults on draft revised slotting guidelines
The FSA begins a round of consultation on its revised guidelines. For the first time, these include some quantitative criteria for categorising loans, such as loan-to-value ratios etc.
Lenders are alarmed and begin to lobby against the revised rules, which are considered to be too prescriptive. They also argue that introducing them in current market conditions would be damaging.
December 2011: FSA backs off introducing revised guidelines
The June 2011 revised guidelines are sidelined. The FSA decides to hold bilateral talks with banks to review their individual positions and adopt slotting on a case-by-case basis.
Jan-current 2012: Slotting is still on the agenda
The FSA undertakes reviews of all banks and other lenders that it supervises. It is understood to be preparing a paper that will propose a way forward, based on its findings.