What drives the performance of Germany's open-ended real estate funds? Daniel Piazolo and Sebastian Gläsner dissect the data

Germany's export-driven economy leads to high outflows of goods and inflows of capital. Some of this capital is then invested abroad through German real estate funds. These German real estate funds can be divided into two main groups: open-ended funds and closed-ended funds. The German open-ended funds account for around €120bn and the German closed-ended funds account for about €68bn. Both groups are invested in Europe and worldwide. However, only the open-ended funds have regular, at least annual, valuations and allow for detailed attribution analysis. Consequently, the following focus is on the open-ended funds.

German open-ended public funds are highly regulated by the German investment law. Those regulations together with industry initiatives to increase transparency have led to a very high disclosure of fund information. For most funds it is possible to perform an attribution analysis including property performance, the effect of financing, liquidity holding and fund costs. The funds can be classified into those that target retail investors and others that target institutional investors. Both groups share the same legal structure: they are open-ended public funds (Offene Immobilien-Publikumsfonds).

The three types of German open-ended funds have the following size: open-ended retail public funds account for €75bn, and the performance analysis is covered by the IPD OFIX publications; open-ended institutional public funds account for €11bn; open-ended special funds account for €32bn. The last two groups are available only to institutional investors and the performance is covered by the IPD special funds study.

The funds that target institutional investors are investment alternatives for institutional money that otherwise would flow into open-ended specialist funds (Offene Immobilien-Spezialfonds). Those Spezialfonds outnumber institutional public funds in terms of number and fund volume, but their reporting is less detailed, and this is why institutional public funds provide the best insights into the build-up of German institutional fund returns.

The sample of institutional funds analysed consists of 15 vehicles with a total net asset value (NAV) of €9.5bn and €11.9bn assets under management which make up about 90% of the market. A big proportion of their assets are invested in Germany (43%) and the rest of Europe (53%), and only a small fraction is located outside Europe (4%). This allocation is related but not identical to the allocation of the much bigger group of public funds for retail investors. Those funds have more assets outside Europe (14%), and less in Germany (28%).

The institutional public funds have outperformed the retail public funds in the period January 2006 to February 2012 by 12.3 percentage points. This equals a return difference of 1.6 percentage points a year.

The outperformance cannot be solely attributed to regional asset allocation differences; because the index of institutional funds outperforms all three subgroups of retail public funds summarized in the IPD OFIX indices, reporting the performance of the open-ended retail public funds. The return development of the OFIX Global index is remarkable, as it led the field of retail funds until mid-2009, but from then on performed much less well. For German open-ended retail funds, negative returns in this dimension are very exceptional and can only in part be explained by the performance of the regional property markets. Some of the funds have entered liquidation recently, and those funds show negative returns, which pull down especially the global index.

The institutional funds can be split into two groups. Group one has a German investment focus, while funds in group two have distributed their assets over Europe. The Europe group consists of eight funds with €6.7bn NAV, and the German group comprises seven funds with €2.6bn NAV. In both groups, some funds are highly leveraged (up to 50%, limited by investment law), while some hardly apply any debt. Even though four funds have stopped investor redemptions, only one fund has entered liquidation.

When accounting for regional asset allocation, the outperformance of the institutional funds persists. Surprisingly, the institutional funds invested in Germany show a more volatile return development than the institutional funds invested across Europe and surpass their returns by about six percentage points over the six-year-period. This is remarkable, because the European investment markets have been more volatile than the German market in this period. Both OFIX sub indices show less volatility, and the OFIX Europe slightly outperforms the German allocated funds, which is in line with higher European property returns than German property returns in this period.

Leverage strongly increased fund performance and at the same time return volatility in the period investigated. On an annualised basis, leveraged funds realised a 4.6% return between January 2006 and February 2012, while unleveraged funds achieved only 3.8%. The outperformance of leveraged funds stems from high returns in the years 2006 to 2008, while in 2009, 2010 and 2011 unleveraged funds take the lead, and outperform the leveraged funds by 1.9 percentage points in 2011.

To further investigate the relation of investor type, asset allocation and leverage to fund return, the figure presents the results of regression analyses. For each year, the model is set up to explain annual fund returns using the factors investor type (private; institutional), region (Europe; Germany), sector (office; other), status (liquidation; no liquidation), and leverage (low; high). The plus and minus signs in the cells indicate the prefix of the variable estimate, indicating the direction of the relation. The asterisks indicate the statistical significance of the relation.

The model is significant for the years 2006, 2008, 2010 and 2011, and between 29% and 48% of the variance can be explained. However, in 2007 and 2009 significance levels are clearly missed, and given the small sample size of 37 funds and their heterogeneous structure, results should be interpreted with caution. The factor ‘investor' misses significance in all six years. In a panel investigation with a higher sample size, however, the factor could become significant, due to the fact that the estimate for private investors is negative for all years with the exception of 2011. A panel investigation was not conducted since the period 2006 to 2011 represents a cycle.

The factor ‘region' is significant only in 2009, and in this period the entire model misses significance. Maybe a more specific analysis of the regional asset allocation would improve results, but due to sample size restrictions this was not possible. The sector allocation, however, is significant in the years 2006 to 2008, indicating that office funds outperformed other funds. The specification ‘other' comprises funds investing in retail properties, residential properties and diversified portfolios, and again a more detailed analysis is not possible given the small number of funds. Even if a regression is not significant for a year this can be deliver valuable insights, since it means that the combination of significant factors for other years does not yield enough explanatory power for that year.

The variable with the highest explanatory power is ‘status'. Funds that have entered liquidation as at March 2012 performed significantly worse in 2011 and 2010, but also already in 2008. As mentioned before, funds in liquidation have a significant negative impact on fund indices. The effect of ‘leverage' is strongly significant in 2008, indicating that low leveraged funds underperformed, and the direction of the estimates (though not significant) fits quite well to the index results. In the boom years 2006-08 leverage had a positive impact, while in times of falling property markets leverage puts a strain on fund performance.

To further analyse how leverage contributes to fund performance, figure  illustrates an attribution analysis of fund returns in 2011. The return components are taken from the annual and semi-annual reports of the institutional public funds, and are being aggregated to capital-weighted index figures. The analysis differentiates between funds with positive leverage effect in 2011 and funds with negative leverage effect, and the return components refer to the left y-axis. The green stacks show the difference per component of both groups, and refer to the right y-axis.

To derive a more accentuated analysis, only funds with high leverage are considered. Out of those nine funds, three funds (€1.4bn NAV) have realised a positive leverage effect in 2011, while for six funds (€3.7bn NAV) using leverage reduced fund returns. The criterion for a positive leverage effect in 2011 was met if a fund reported a higher return before leverage  than after leverage.

The first five return components from gross income to total return represent the performance on the direct property level. With regards to gross income, operating costs and net income, the funds that achieved a positive leverage effect performed worse than the others. But capital growth makes the difference, as the properties of the one group increase in market value by 1.3%, while the other groups' properties lose 1.8%. The positive leverage effect group outperforms the other group by 3.1 percentage points in this return component. Foreign taxes have only minor effects on both groups, but the difference in returns before and after leverage illustrates the consequence of using debt in a situation of increasing property values versus decreasing values. In the first case, the property return before leverage increases from 5.4% to 6.8% after leverage, while in the second case the return before leverage decreases from 2.7% to 1.7%. The return differences of both groups that mainly stem from a different capital growth aggravate from 2.7 percentage points to 5.1 percentage points. All other things being equal, the use of debt strongly increases the return volatility.

Currency also has a minor effect on the fund returns, but given a predominant European asset allocation, this does not surprise. Liquidity returns are at approximately 0.6%, with very small differences between the groups and also between the individual funds. The total expense ratios are at 0.8% (negative leverage effect) and 1.1% for the funds with positive leverage effect.

After all amplifying and dampening effects of the fund structure, the total returns on the direct property level of both groups were higher than the final fund returns. This is in line with the fact that leveraged funds outperformed from 2006 to 2008, but underperformed in 2009, 2010 and 2011 compared with low-leveraged funds. Direct property returns of the positive leverage effect group of 5.6% were transformed into 4.9% fund returns. The direct results of the negative leverage group decreased from 3% to 1.2% on the fund level. The return differences of both groups increased from 2.5 percentage points on the direct level to 3.7 percentage points on the fund level.

In the year 2011 and for the given sample of property funds, the use of debt increased return differences stemming from different capital growth developments, and were slightly beneficial for the one group, but harmful for the funds with a negative leverage effect. In times of strongly rising markets, the results will be much more in favour of high leverage, but the analysis illustrates empirically that using debt strongly increases fund return volatility.

Most, if not all, managers are convinced that they do a better job than the average. With their expertise and resources, institutional fund managers can be expected to do a good job. Attribution analyses are important tools to determine why some fund mangers are able to deliver an outperformance and can live up to their own convictions and investor expectations.

Daniel Piazolo is managing director and Sebastian Gläsner is funds services manager at Investment Property Databank (IPD)