With the risk-free rate effectively zero, what is a sustainable loan-to-value ratio for investors? Gerry Blundell looks for some answers

There can be few investors who still see even UK government debt as the risk-free rate. Systemic risk has been passed from equity markets (2000-02) to the private sector (sub-prime), finally landing in the laps of government as they bailed out the banks. Today, risk-free rates are effectively zero or slightly below; some investors consider even the most volatile commodity, gold, to be a safe haven.

Efficient market theory argues that lower risk-free rates push the optimal portfolio towards cash and bonds, and this trend has been apparent in institutional portfolios. A combination of shortening liabilities and proposed regulations such as Solvency II is forcing institutional funds towards long income, especially index-linked. The implications of the new normal for real estate are profound: UK real estate assets that can offer a secure stream of income are trading at yield premia of more than 200bps below those that cannot.

It is ironic that as theory and regulation push UK institutions away from equity toward bonds, suitable real estate-backed debt is near unavailable. In December, De Montfort University reported that only about £12bn of debt had been refinanced in the first half of 2011 and that £85-114bn could not be refinanced on current terms.

Yet, while zero remains the new normal, fixed income secured against real estate should be attractive to investors, arguably more so than many governments' debt. So, assuming that sooner or later existing debt will have to be re-structured, this raises the question of what terms are likely to prove sustainable to both lenders and borrowers over the longer term, given the vicissitudes of the UK markets.

What is a sustainable loan-to-value (LTV) ratio? Looking back over past trends in real estate values one may observe how frequently loans at different LTVs would have come to grief. Data on this are available at an aggregate level back to 1971 from IPD and before then from Paul Scott (‘The Property Masters') back to 1921; excluding the second world war, a period of 90 years. Because these data are based on the aggregate performance of several properties, results should be regarded as more appropriate to portfolios than single-asset loans.

Looking at all the five-year periods that would correspond to a typical loan, values fell in a quarter of cases but never by more than 30%, so it might be presumed that an LTV of 70% was safe, at least from the lender's viewpoint. Moreover the 70% LTV covenant would only have been breached once during those loans, in the two, three and four-year periods ending in 2009. What was different about the run-up to 2009 was that inflation was well below the 90-year average: in the five years to 2009 it averaged 2.8%. compared with 3.6% since 1921. A small difference but enough to make a big impact.
If one strips inflation out of the long-run data, capital returns fall from a nominal 3.1% pa to a real decline of -0.5% pa, and the chances of a five-year loan at 70% LTV ending up under water rise to 16% of the five-year periods. Longer-term loans fare worse because the long-term trend in real capital value has been slightly negative. Over the whole period, a 10-year bullet loan at 70% LTV has an 18% chance of seeing values fall by more than 30%, and a 20-year loan 41%.

The moral is that, with zero inflation, longer-term loans either need to be in the lower LTV space to offset the increased chance of failure or parties to the loan need to be consistently good at timing the market - a tall order.

It can be argued that, looking forward, assuming zero inflation is unrealistically pessimistic. Historically real estate capital returns over five years exhibit a +0.54 correlation with RPI, so it is likely that future inflation will be partly reflected at least in capital returns. Therefore, the LTV risk analysis was reworked assuming 2.5% pa inflation, the rate implied by the UK gilts market at the time of drafting.

The presence of inflation radically changes the shape of LTV risk, with longer-tenor loans being slightly less risky than five-year debt. For loans at 70% LTV, the bottom row is the key one, suggesting a 4% chance of a 30% or more fall in asset value over five years, 2% over 10 years, etc. At 2.5% inflation, the analysis suggests portfolio loans at LTVs of less than 60% are secure, at least based on the evidence of the last 90 years.

This conclusion comes with a caveat. It assumes the past is a good predictor of the future. However, data with that mean reversion are usually normally distributed. This is not the case with property values over five-year periods. The dispersion is skewed to the left, towards losses in value, and shows evidence of twin peaks. Similar dispersions can be observed in more recent data samples.

A practical solution to this would be to build in a comfort margin of, say, a maximum of 50% LTV on portfolio-backed debt, and even lower levels on individual assets. In a 2011 Investment Property Forum study into the factors driving property portfolio risk, it was found that risk (measured in terms of tracking error) started to rise exponentially above 30-40% LTV, pointing to this level being a critical risk threshold.

While 30-40% LTV may be sustainable over the long run, it presents serious issues for refinancing existing debt as it implies a huge level of refinancing at a time when equity exposure to real estate is threatened by changes in regulation. It has been estimated that £25bn of existing debt is at an LTV of more than 100% as values have fallen so far since the debt was originated. Cutting this to 40% would require some £15bn of fresh equity: equivalent to over three years' net investment into the IPD universe.

What is needed instead is greater variety in the terms of loans and the sources of debt. Self-amortising loans would help spread the shock of lowering LTV levels over several years; and they would imply relatively modest reductions in returns, especially while capital growth expectations are muted. A 50% bullet loan at 4.6% interest on a 7% yielding asset produces a leveraged return of 9.4%, even with no capital growth. The same terms with amortisation down to 30% only reduces returns to just over 8% pa for a five-year period. As investors' return expectations deflate, the pressure for excessively high rates of gearing should diminish too, at least until the next triumph of hope over experience.

Looking forward, it is possible we shall see finance originate from a wider range of sources than before. A combination of Solvency II and the sidelining of banks may well lead to savings institutions increasing their senior debt exposure at the expense of direct ownership, and with longer tenors more matched to the length of their liabilities. The higher-risk equity stub and the more junior tranches of debt attached to the asset may then prove to be attractive to private equity and the CMBS market, which will bring a range of debt/equity hybrid instruments to the asset class.

One thing is certain: for UK commercial real estate: the debt genie is out of the bottle. The cost of morphing the genie into a more sustainable form will eventually be shared between the banks' haircuts, investors reducing return expectations, and fresh sources of risk equity.

Gerry Blundell is strategic adviser at Legal & General Property