Is there a relationship between REIT prices and pricing in the growing market for secondary trading in unlisted funds? Paul Schneider investigates

Secondary market exchanges of UK unlisted real estate fund interests have evolved significantly since the 1990s, when only a few fund managers offered matched-bargain services. Billions of pounds of fund interests are traded each year by multi-managers and specialist brokers. Increased liquidity in the secondary market allows investors to better manage exposures and potentially reduce buy-sell spreads and avoid ‘J-curve’ establishment costs. 

Unlisted funds hold heterogeneous private real estate assets, but their structure facilitates ownership of homogenous divisible units. Holdings in these funds entitle investors access to a cash-flow stream that is derived from the underlying property assets. Estimations of the market values of such holdings have historically relied on net asset value (NAV) – a backward-looking view of assets, less an independent view of liabilities. Fund assets (comparables for which themselves may be traded reasonably frequently) are valued by independent valuers or appraisers, and these valuations subsequently form part of the fund NAV, from which a fund’s NAV per unit is calculated. 

The Royal Institution of Chartered Surveyors (RICS) Red Book provides guidance on the open market value of property but does not detail how fund interests should be valued. In a liquid market of recent, relevant transactions, the RICS approach serves well to avoid inflationary bias while still capturing pricing of current or near-current pricing. 

However, in trending markets, the inherently backward-looking nature of comparison-based valuations hinder NAVs from adequately reflecting the perceived spot market. Volatile markets – where underlying asset values decoupled from NAVs and, consequently, open-ended fund managers were unable to uphold liquidity provisions – have heightened scrutiny over unlisted fund valuations. 

Fund structure can distort the underlying asset cash-flow profile to the extent that the fair value1 of an investor’s fund interest differs from their units’ NAV per unit. Importantly, the risk that investors ascribe to these distortions, which help in forming their view of fair value, might not be adequately captured by a fund-unit NAV. A pertinent question for funds is whether it is appropriate to value fund units just as a proportion of the underlying assets and liabilities without giving regard to the factors that might distort the investor’s cash-flow profile.

Pricing in the secondary market might represent the reconciliation of NAVs to what the investor deems fair value of his equity. If pricing is simply a correction for a stale NAV, then it may not be providing any forward-looking information. On balance, however, by removing the anchoring to NAV and noting the dominance of sophisticated investors in the secondary market, it is expected that this market provides information relevant to price discovery.

Premia/discounts (price/NAV) of 526 secondary transactions2 of UK unlisted real estate funds between February 2007 and May 2013 were examined. 

The first objective was to examine the relationship between secondary market pricing and the UK’s publicly listed real estate investment trust (REITs) market. Secondly, analysis was performed to better understand which factors (fund and non-fund specific) investors use to price their interests differently from NAV. Any such findings may contribute to the debate as to how the NAV methodology can move towards more investor-relevant fair value-type metrics. Lastly, analysis was performed to determine whether the UK secondary unlisted-funds market contained any price discovery (predictability of future fund returns), which is incremental to information from REITs.

Figure 1 shows the strong relationship (correlation = 0.74) between the monthly weighted average of secondary market trades and past (four-month lag) REIT pricing. 

1. Global real estate multi-asset capital value indices

Although REITs and the unlisted market price in similar patterns, they are certainly not the same. Despite the much lower trading volume, the unlisted funds market consistently trades at lower amplitudes away from NAV (3.14% discount) than REITs (15.08% discount). Hence, pricing in the secondary funds market must contain information that is separate from the REITs market. This separate (potentially additional) information that is unique to the secondary market warrants investigation of certain factors (both fund and non-fund specific) that are associated with fund pricing.  

Factors regarding fund pricing
A panel regression model of fund secondary pricing was run against fund and non-fund specific factors to examine their association (figure 2).  

2. Fund secondaray pricing vs fund and non-fund specific factors

• Non-fund-specific factors: Past REIT pricing for the sector of the transacted unlisted fund is significant in the pricing of secondary unlisted fund interests, a result consistent with figure 1. Inclusion of REIT pricing is an important control variable for the examination of other factors. 

Secondary market trade volume is not statistically associated with fund pricing. As the secondary market expands, one might expect the relationship between volume and pricing to strengthen. BBB credit spreads were significant and act as another control variable in isolating the relationship between fund pricing and fund characteristics.

• Fund factors – past returns: Higher past (six month) fund returns lead to a higher (lower) premium (discount) which suggests investors are extrapolating past returns. This could be interpreted as a violation of the ‘weak form market efficiency hypothesis’. However, it is also possible that in a relatively illiquid market with barriers to entry, investors are using (correctly or incorrectly) past returns as a blunt proxy for management or fund quality. Fund yield was shown to be positively related to secondary fund pricing. This effect is likely to be similar to past returns but also could reflect investors’ willingness to pay a premium for secure income in the recent low yield environment. 

• Size: Fund size is not associated with secondary market pricing. This is surprising, given evidence of a positive relationship previously documented in the public market but also the apparent importance of transparency in less developed markets. Discussions with brokers of unlisted funds prompted a subsequent examination by fund structure. It was found that open-ended funds (which are the majority of the sample) are 60% larger than closed-ended funds, but due to their liquidity mechanism most open fund trades are completed close to the manager’s bid-offer spread. To move outside this range requires a considerable redemption (or subscription) queue. Hence, open-ended funds have less scope to produce higher premiums due to the likely bid-offer bound. 

• Diversification: Asset diversification was found to be insignificant. This is a surprising result if one subscribes to the theory that fund investors value diversification. However, discussions with brokers revealed that factors of management quality, asset performance and distribution yield are the dominant demand drivers. Also, many of the secondary-market investors are multi-managers who are forming portfolios of different real estate exposures – hence they are not necessarily valuing diversification at the fund level, but rather create it themselves by acquiring a range of different fund interests.  

In contrast to asset diversification, income concentration is positively associated with pricing, the opposite of what diversification predicts. The positive relationship may be linked with arguments that diversification is not valued, in fact penalised. However, in this case, it is more likely the variable has inadvertently picked up a credit-quality effect in that tenants who have the capacity to pay larger amounts of rent are likely to be of higher credit worthiness. 

• Leverage: Leverage does not lead to higher discounts. One could have reasonably expected leverage to be the most important example of a risk factor that affects investor cash-flow profile, but it is not accounted for by NAV and hence should be relevant in secondary pricing. However, there is no evidence to suggest leverage is associated with different levels of secondary-market pricing. It was posited that, because leverage adds risk that is not captured by NAV methodology, more highly levered funds would price less favourably. However, the insignificance of leverage as a factor might suggest leverage is only penalised by investors in trending markets, when NAV and fair value decouple at an accelerating rate.

• Investor base: Funds with a greater concentration of investors were priced less favourably than funds with a more diversified investor base. Thus, it is likely investor concentration has been a suitable proxy for liquidity and is consistent with the priori that less liquid funds price worse. This pricing of fund illiquidity highlights the difference between NAV and the fair value that investors ascribe to their units.

• NAV staleness: NAV staleness was not statistically significant in explaining secondary market pricing. It is possible there is noise in this proxy given that it is difficult to pinpoint the date (possibly due to drawn out deal processes) at which the pricing reflects the investors’ views of fair value.  

• Price discovery: If pricing is simply a reconciliation of NAVs to what the investor deems fair value, then it may not be providing any forward-looking information. On balance, however, removing the anchoring to NAV and noting the dominance of sophisticated investors, it is expected that secondary market prices provide information relevant to price discovery.

Barnes and Robinson (2012)3 suggest that “Fair Value for a fund interest should typically lead the trend in NAV as investors appreciate where the spot market is and anticipate the evolution of property valuations.” 

For example, if secondary market investors price a fund interest at a substantial discount, they may be taking a view that the fair value of the fund interest is not as high as its NAV. If they are correct, NAV should subsequently fall, which will be reflected in the lower total return of the fund. 

Using a panel regression model, future (unrealised, NAV-based) returns are regressed against current secondary fund pricing (proxy for price discovery) and a number of control variables that were previously identified as significant in current pricing (figure 3).  

3. Future returns vs secondary pricing and control variables

The key result is that current fund premium is positively associated with future (unrealised) returns. As investors price secondary market interests above (below) NAV, they are taking a view that NAV under or over-prices fair value. If investors are correct in their assessment, subsequent changes in NAV should be higher (lower). This result is opposite to studies of REIT discounts and future returns where larger discounts are associated with higher subsequent returns. I

n these REIT studies the return is more or less the investor capturing the subsequent stock price gravitating back to NAV – hence the discount, in part, becomes the return. This study, however, in the absence of frequent repeat fund transactions, uses (unrealised, non-transaction) returns, as calculated by IPD. Using unrealised returns that does not incorporate the discount means variability in total returns is essentially NAVt+6NAVt , given that income is relatively stable. This more stringent definition of total returns means the predictive power of current pricing for future realised returns, should they occur, is understated, given the sample has a bias to discounts.    

Critically, the significance of unlisted fund pricing remains significant in the presence of listed market pricing. This suggests that the secondary market has explanatory power for future (unrealised) returns (changes in NAV) that is separate and additional to the information that investors may absorb from the REIT market. 

This study produces the first documented evidence that UK unlisted real estate fund interests are related to past REIT pricing. While similar in movement, they are different in magnitude. In addition to unlisted funds’ relationship to REITs, funds’ specific factors of past returns, yield, income concentration and ownership concentration are associated with pricing. 

Pricing in the secondary market seems to be a combination of macro and fund-specific factors that are potentially under-priced by NAV. The secondary market’s ability to predict future unrealised gains (approximately future changes in NAV) was in addition to information flowing from REITs and robust to fund structure.  

This study’s results suggest investors can use the secondary market for guidance on the fair value of their fund interests. It should be noted, however, that the benefit of increased transparency is likely to coincide with pressure from auditors for investors to value fund interests more often. Nonetheless, even with the higher volatility that this would induce, increased transparency should help the primary funds market raise more capital as investors gain comfort in the knowledge of a liquid secondary market. 

While this study has endeavoured to take steps to reduce any potential bias, some caveats remain. The study was not able to capture a full property cycle, nor was it able to capture the pricing that occurs between private investors. Although fund premium was, on average, only at a slight discount, the skew in the series was still large and may have econometric consequences, even after data was winsorised.

There is a delay between when deal due diligence is commenced and when the transaction is time stamped. Hence, it is possible that the pricing of the deal is not reflective of investors’ view of fair value at the time indicated by the transaction data. The insignificance of NAVAGE (number of days between date of NAV to which deal was underwritten and trade date) supports this view. 

As the secondary market develops it will provide researchers with more detailed data allowing them to examine secondary market pricing over a full property cycle and to examine the impact of fees, bid-offer spreads and distribution policies on fair value. 

Footnotes:

1 IFRS 13: the price that would be received to sell an asset, or paid to transfer a liability, in an orderly transaction between market participants at the measurement date

2 Source: CBRE-GFI PropertyMatch and JLL

3 Barnes, G. and P. Robinson. (2012). ‘Fair enough: getting fair value right – it’s not all about NAV’. CBRE research note  

Paul Schneider is associate portfolio manager at QIC. Alex Moss is managing director at Consilia Capital

In favour of global diversity