The above analysis was restricted to a national context, as its purpose was to access the impact of constraining real estate investment on risk-adjusted performance. However, as discussed in section 2.3.2, the relaxation of UK exchange controls in 1979 is often highlighted as the principal reason why life and pension funds limit their exposure to real estate.
As table 3.14 below illustrates, allowing investors access to Far East equities reduces their exposure to real estate. However, these results are qualified in that highly risk-averse investors reduce their investment in UK equities to invest overseas, while less risk-averse strategies reduce their investment in real estate.
Nevertheless, when compared with current life and pension fund practice, the model still proposes that a substantial proportion of portfolios be held in real estate. Theoretically, if the investment universe was further expanded, then the proportion held in real estate would be further reduced. However, the model is unlikely to provide evidence consistent with current life and pension fund practice.
Explanation for the inconsistency observed may alternatively be sought in the further unsmoothing of the real estate returns series. However, in order to reduce real estate's average allocation to 10% (i. e. twice the current average life and pension fund exposure) the variance of the capital returns series would have to be increased by over 280%. This is approximately three times that achieved using a=0.5, the upper bound proposed by Barkham & Geltner. In addition, it has been argued that the impact of appraisal smoothing is likely to be less important in lower frequency returns series. 25 However, by employing annual data as in the above study, the strength of this argument is limited.
25See for example, Giliberto , Liu, Hartzell, Grissom and Greig  and Fisher, Geltner and Webb .