En intressant artikel om index derivat och lån som istället för fast ränta följer ett bostadsprisindex är "Property Derivatives and Index-Linked Mortgages" av Juerg Syz & Paolo Vanini & Marco Salvi publiserad den 15 augusti 2007.
Här följer ett utdrag ur artikeln:
Economists have forcefully argued for the introduction and use of property derivatives as a hedge against house price risk (e.g. Shiller and Weiss, J. Real Estate Finance Econ., 19(1):21–47, 1999). The rationale for these financial instruments seems clear, as many households are heavily invested in housing and standard financial instruments offer a poor hedge. In practice, however, most of the property derivatives available have been targeted to meet the needs of institutional investors, not those of owner-occupiers. Building on the recent launch of the first Swiss property derivative, we here propose index-linked mortgages tailored to retail consumers. The payments of these mortgages depend on the corresponding housing market performance. We further price the instruments, discuss the stabilization of the homeowner’s net wealth, and quantify the expected decrease in the mortgage default risk achieved by this immunization effect.
Is an Index-Hedge Appropriate? The price of housing is subject to considerable fluctuations over time, which in turn lead to significant fluctuations in wealth. As pointed out by Sinai and Souleles (2003), the effects of house price risk on consumers’ choices are ambiguous. For a household with utility defined over housing consumption, homeownership acts as a hedge against changes in the cost of consumption, i.e. against rent risk. Housing market risk may thus increase homeownership rates. The extent to which hedging considerations affect tenure choice is mitigated by the existence of frictions in the real estate and mortgage market. Transaction costs coupled with borrowing constraints restrict the number of house trades and investors’ ability to implement first-best strategies significantly (Cocco 2000). In this context, the existence of a liquid property derivative, such as the one described, is of great interest to investors. However, the extent to which an investor may take advantage of such a market is limited by its effectiveness for hedging purposes, i.e. by the amount of the idiosyncratic risk of individual properties compared to the risk of the overall market.
Property Derivatives and Index-Linked Mortgages
Results from previous research indicate substantially higher variability in returns of individual properties relative to the ones of the national market. For the Swedish market, Englund et al. (2002) report a standard deviation of the returns of individual properties of 11.3%, compared with 7.6% for the market as whole. Goetzmann (1993) also documents a substantially higher variability for individual properties than for the regional market, with standard deviations 1.5 to 3 times higher for four US metropolitan areas. For New Zealand, Bourassa et al. (2005) find standard deviations 1.4 to 2 times higher than the ones of the general market. They relate the degree of variation in price changes among houses within a market to their characteristics and to the prevailing conditions of the housing market at the time of the sale. Atypical houses and houses with characteristics in limited supply, for example waterfront houses, are generally more risky. Unfortunately, the Zurich residential market is quite illiquid, with less than 2% of the housing stock traded each year. Even though our database covers around 20% of all transactions, we were able to identify only 264 repeated sales out of a total of 14,000 transactions. In our limited sample we measured a low correlation (0.19) between individual and market returns over a horizon of 4 years, rising to 0.38 over an interval of 8 years, and to 0.68 over 12 years. Although based on a very limited sample, these results are in line with the work of Iacoviello and Ortalo-Magné (2003) for London, who find a weakly positive correlation of 0.13 between the London housing returns and simulated individual returns at a short horizon (1 quarter) but a very strong correlation (0.87) at a 10-year horizon. In order to check for the robustness of our results we implement their methodology, adapted from Englund et al. (2002).
We take advantage of the availability of 11 geographically disaggregated indices for the Zurich area to approximate the returns on individual properties. We thus assume that the idiosyncratic variation of individual housing returns is captured by the variation of the local index returns around the returns of the market index. Note that this method is likely to underestimate true idiosyncratic risk, as the regional segmentation only represents one source of idiosyncratic risk. The estimated idiosyncratic volatility of the returns σv is equal to 0.082 while the volatility of the yearly index returns σI is 0.054. The correlations are indeed very high, ranging from 0.77 for 4-year, to 0.89 for 8-year and 0.96 for 12-year periods. In short, these partial and preliminary results confirm the importance of designing hedging instruments with maturities of at least 5 years, as correlations are higher at longer horizons.