Optimal portfolio allocations to UK commercial real estate

Introduction

As demonstrated in our earlier research, structural shifts in how people and organisations connect, consume and work have led different real estate sectors to exhibit divergent performance in recent years. For example, retail has been challenged by the growth in online shopping over time while logistics has benefitted from the same trend. The office sector has faced challenges from the adoption of hybrid working models. Such trends have posed challenges for strategic asset allocation – the process of deciding which asset classes or types of assets should receive investment moving forward. This normally precedes the selection of individual assets.

One approach to asset allocation could be to simply invest in asset classes and asset types that are expected to provide the highest returns. However, high returns might only be achievable if the investor accepts greater variability in performance outcomes, or another drawback such as reduced liquidity. Therefore, many investors will seek to balance potential returns against the risks being taken to achieve them. One way to reduce risk is to diversify across a range of investment opportunities, raising the question of what the optimal combination of real estate asset types has been and what this might be in the future.

We explore this topic within the context of UK commercial real estate and its main sectors, reflecting on developments over the past twenty years. A balanced real estate portfolio has traditionally consisted of investments in office, retail, and industrial properties. However, because of the structural changes outlined above, the share of capital allocated to ‘traditional’ real estate sectors has changed. Investors have broadened their portfolios in recent years to embrace a range of ‘alternative’ asset types, including investments into the residential, hotel, and healthcare sectors.

We examine the extent to which these changes reflect movements by real estate investors towards an optimal mix of segments as implied by an analysis of property performance data. We analyse the ten years to 2013 and the ten years to 2023 to see how the optimal combination in a portfolio setting has changed. Understanding these changes is crucial for investors who own land and property directly, as well as those who gain exposure to real estate via investment in private property funds or REITs. In the latter case, portfolio construction involves selecting funds or REITs with assets and strategies that align with the investor’s strategy and performance objectives.

Method

We apply Modern Portfolio Theory and calculate optimal portfolios for commercial real estate using data from the CBRE UK Monthly Index. The portfolios presented are based on historical total returns for the office, retail, industrial, and other sectors. ‘Other’ includes investments in asset types such as leisure, hotel, student accommodation, and self storage. Although we do not examine how these different segments might perform in the future, the analysis illustrates a base model for allocating capital in real estate and it could be applied using views on target sectors moving forward.

We annualise the monthly total return rates recorded for each segment and Then calculate the mean, standard deviation, and covariance between segment return rates to determine a minimum variance frontier. This frontier is a series of points (or portfolios) where the combination of investments chosen delivers the lowest possible standard deviation for a given level of return. We also identify the optimal portfolio as that which provided the best performance relative to the variability of its returns as captured by the Sharpe Ratio, a common measure of risk-adjusted return.

We analyse the data in nominal and real (inflation-adjusted) terms, recognising that some investors prefer to invest in assets whose returns more closely track movements in inflation. Meanwhile, optimisation models can generate counter-intuitive solutions that do not mirror how investors approach portfolio construction. Therefore, we imposed certain limitations in our analysis: no single sector (such as office) could exceed 50% and no single segment (e.g. Central London office) could exceed 30% of the investment allocation. We also specified that a maximum of 10% could be allocated to 'other' sectors to reflect the challenges associated with large-scale investments in these asset types during the periods reviewed.

Figure 1: Average annualised returns and standard deviations for UK commercial real estate segments

Ten years to December 2013

UK real estate performance during the ten years to December 2013 was significantly impacted by the Global Financial Crisis (GFC). All property capital values declined by 44% from July 2007 until July 2009. While real estate investors did benefit from strong returns outside of this two-year period, the decline in capital values during the GFC heavily impact both overall return and volatility.

In this period, all property had an average annualised total return of 7.7% with a standard deviation of 15.8%. Industrial recorded the highest average return, at 11.5%, followed by Central London offices (9.5%), but these segments had high standard deviations as well. The ‘other’ segment offered the best risk-adjusted return of the individual segments, perhaps reflecting the variety of assets included in this grouping.

Figure 2 shows the minimum variance frontier and optimal portfolio based on nominal return rates. For the ten years to 2013, the optimal allocation would have been 50% to industrials, 30% to Central London offices, and 10% in each of Outer London/M25 offices and ‘other’. This portfolio had an average return of 10.1% per annum and a standard deviation of 16.4%. Although the average return was lower compared to the industrial sector alone, it was offset by reduced volatility in return over this period.

The optimal portfolio did not contain any allocation to retail segments, but other portfolios at lower levels of risk and return did include the retail sector. However, retail plays a larger role when analysing real returns. In this case, the optimal portfolio comprises 30% in each of Industrials and Central London offices, 21% to Retail Warehouses, 10% to Other, and 9% to Outer London/M25 offices. As such, it is closer to the actual allocations adopted by multi-sector real estate investors in this period.

Figure 2: Optimal real estate portfolios for the ten years to December 2013 

Ten years to December 2023

UK real estate performance in the ten years to December 2023 was influenced by both cyclical and structural factors. The rise of online shopping and hybrid working led to a divergence in investment performance across sectors, a trend that was further accelerated by the effects of the COVID-19 pandemic. This is illustrated by the performance of some segments relative to the previous decade. For example, Central London office average returns declined from 9.5% to 6.6%, while the average return for the industrial sector rose from 11.5% to 15.3%.

The performance of all segments was also affected directly by the COVID-19 pandemic, and then by the rapid and significant rise in UK interest rates to combat inflation. Increases in both short and long-term interest rates triggered an outward movement in real estate yields. This caused all property capital values to decline by 22% between June 2022 and December 2023.

Figure 3 shows the minimum variance frontier and optimal portfolio for the ten years to 2023 based on nominal returns. For this period, the optimal portfolio has an average return of 11.2% and a standard deviation of 16.2%. The allocation is similar at sector level to the previous ten years, but the combination of segments in the office sector differs. 50% of capital is allocated to the industrial sector, with 30% allocated to Rest of UK offices, and 10% each to Central London offices and the ‘other’ segment.

Once again, the optimal portfolio did not contain any allocation to retail, but other portfolios at lower levels of risk and return did include the retail sector. The optimal portfolio based on real returns allocated 30% each to Industrials and Rest of UK offices, 20% to Central London offices, and 10% each to ‘other’ and retail warehouses. Overall, real returns were impacted by the high inflation since 2021. Nonetheless, the average real total return of the optimal portfolio was still 5.2%.

Figure 3: Optimal real estate portfolios for the ten years to December 2023

Comparing the results

There are some similarities in the results from each period. This includes the large allocations to the industrial sector in the optimal portfolios, the greater role for retail warehouses when working with real returns, and the fact that the ‘other’ segment always received an allocation up to our imposed maximum. Results incorporated the impact of a major market downturn in each period, with the GFC having an even greater effect on returns and volatility than the downturn in June 2022.

However, a key difference is apparent when comparing the minimum variance frontiers for each period. This comparison shows that similar levels of total return were reached at lower levels of standard deviation for the ten years to December 2023, leading to higher risk-adjusted returns for portfolios on this frontier. Moreover, the optimal portfolio for the later ten-year period had a higher average annualised return (at 11.2% compared with 10.1%) for a slightly lower standard deviation in returns (16.2% vs. 16.4%).

This outcome might seem surprising given the multiple challenges faced by the UK real estate market in recent years. However, it stems from the divergent performance across different property types. This has reduced the covariance between segments and has been beneficial from diversification viewpoint. Still, careful selection of individual investments is needed to avoid low returns in sectors that are undergoing structural change.

Do these results mirror investment in practice?

Large allocations to the industrial and ‘other’ sectors is also reflected in the MSCI UK annual index. The industrial sector’s share of the index’s total capital value increased from 15% in 2013 to 32% in 2023. Similarly, the ‘other’ sector increased from 6% to 11%, and residential from 5% and 9%. This highlights the shift away from the office and retail sectors. Similarly, UK investment volumes have transitioned toward the industrial and residential sectors.

Industrial’s share of annual transaction volumes increased from 10% in 2013 to 19% in 2023, while the residential sector has seen a significant rise in investment, with its share increasing from 3% to 25% over the same period.

The segments in our analysis were largely based on property types, but this represents only one approach to achieving diversification in a private equity real estate portfolio. Other approaches can include more detailed treatments of location and other asset attributes such as size or lease length. It is also possible to de-smooth property performance data if there are concerns that volatility is understated. However, we have not taken this step in this instance as we do not compare real estate returns to those from other, publicly-traded asset classes such as equities and bonds.

Historical analysis of real estate performance can inform the strategic asset allocation process, providing a basis for estimates of volatility and correlations between different market segments. Forecasts of expected returns and views on how volatility and correlations might change are also important, as these inform how allocations should be changed in future. However, our historical analysis suggests that all commercial real estate sectors could contribute to the construction of effective portfolios, especially where the volatility of both nominal and real investment returns matters.