U.K. Defined Benefit Wind Down: Sizing the Risks and Opportunities

An exercise in combining disparate data sets

Arriving at an estimate for the composition of the U.K. DB real estate portfolio is not an exact science and requires working across datasets that may not always be entirely overlapping; there is, therefore, a margin of error around our conclusions but we are confident they are approximately correct (rather than precisely wrong). Figure 1 sets out the sources that we used to arrive at our numbers.

• First, we began with ONS data, which told us the total size of real estate holdings held by funded occupational pension schemes in the U.K. The data contains figures for public DB, private DB and private defined contribution (DC) (though we are not concerned with the latter here) and also provides figures on the proportion held by each in direct and pooled strategies.
• Our second data source was MSCI. We assumed the U.K. DB direct real estate portfolio to have the same segment weightings as the MSCI Annual Index. For the U.K. DB pooled real estate portfolio, we assumed that the weighting of conventional versus long income is as per the MSCI/AREF Property Fund Index (roughly 75:25). We then assumed that the segment weightings for each part of the U.K. DB pooled real estate portfolio are the same as the corresponding part for the MSCI/AREF Property Fund Index.
• We then turned to our third data source, the CBRE Long Income Index. Since we wanted to focus purely on conventionally leased real estate, we wanted to exclude income strip and ground rent structures. The CBRE Long Income Index shows the proportionate value of these in their long-income fund sample. We assumed that the long-income exposure within the U.K. DB pooled real estate portfolio was similarly weighted.
• Finally, our fourth data source was RCA. This data allowed us to understand historical average liquidity volumes at the segment level and investor composition. By excluding a proportion of the investment potentially accounted for by DB money, we arrived at a conservative view of potential future typical liquidity against which we could benchmark the U.K. DB real estate portfolio.

Figure 1: Estimating the DB portfolio: sources and methodology

Splitting out the £78 billion DB real estate portfolio

The results of this exercise are shown in Figure 2 where we break out the private DB and public DB real estate portfolio segment weightings.

• In total, the U.K. DB real estate portfolio is approximately £78 billion. Private DB accounts for around £42.6 billion of this and public DB for around £35.6 billion.
• Within private DB, £22.0 billion of the real estate portfolio is direct (52%), £15.1 billion pooled conventional (35%) and £5.5 billion pooled long income (13%).
• In public DB, the largest exposure is to pooled conventional at £16.6 billion (47%), rather than direct at £13 billion (37%). Long income remains the smallest component accounting for £6 billion (17%).

Figure 2: Estimated size and composition of DB real estate portfolios

Understanding movement in the DB portfolio

We have already used the ONS data to understand the current size of DB real estate allocations, but that same data allows us to quantify how that has changed. Figure 3 compares the portfolio size at the peak of the market, in June 2022, with the latest available datapoint, September 2024. Over this timeframe, the value of public DB property holdings decreased by £5.6 billion to £35.6 billion, and private DB decreased by £23.7 billion to £42.6 billion. Meanwhile, public DC holdings have increased by £5.2 billion to £9.5 billion.

During the same period, MSCI all property values declined by -22.4%. Assuming that DB and DC property portfolios performed in line with this market average, a significant amount of the above value change would be attributable to market movement.

• For private DB, the story is roughly of three parts value fall and two parts disinvestment. Of the £23.7 billion fall in the value of private DB property holdings, -£14.9 billion of this would be attributable to market movement, implying a further -£8.8 billion of net sales of property.
• Public DB has apparently been a net buyer over the period since the -£5.6 billion decline in the value of the property holdings is greater than the -£9.2 billion decline that would arise from market movement. The balance of +£3.7 billion is the implied value of net purchases.

Figure 3: Reconciling the change in value of DB and DC holdings

We are concerned in this report with DB holdings of real estate that are likely to be targeted for disposition if not already, then in the near future. The (attempted) sale set of those assets will offer opportunity to investors seeking to buy from motivated sellers and risk to the wider market should such dispositions become disorderly. The above analysis shows us that for our purposes, we need only consider private DB real estate holdings. While public DC holdings will see transaction activity and perhaps a significant change in structure weightings, as pooling encourages greater direct ownership, the direction of travel of the whole will be to increase means, which are less relevant as a source of shorter-term risk and opportunity than private DB.

Raw values are useful, but comparison against typical liquidity is also meaningful

Our goals for this analysis are to show where market risk might lie in the case of a rapid and/or disorderly sell-down of the private DB portfolio, and assessing the scale of opportunity for investors looking to acquire assets from a motivated seller. For the latter purpose, the left hand charts in Figure 4 and Figure 5, showing the size of the private DB portfolio in value terms, is ideal. To understand the potential for market dislocation, the right hand charts, showing private DB holdings in terms of typical market liquidity, are more instructive.

For the £40.5 billion conventional (i.e., excluding income strip and ground rent) private DB real estate portfolio, we estimate:

• Standard industrial accounts for approximately £10 billion (around three quarters of the total private DB real estate portfolio), with the majority in the South East. A further £5 billion is in distribution warehouses, roughly equally split between the South East and Rest of U.K.
• Office accounts for a further £8 billion, with the majority, £5 billion in London.
• Retail exposure is at £7 billion; almost half in retail warehouses and a further 20% in supermarkets. Half of the exposure to supermarkets is in the long-income allocation.
• The remaining £10 billion is accounted for by £4 billion of residential (more than half of which is PBSA), £4 billion other and £2 billion hotels.

Collectively, the £40.5 billion exposure represents 0.9 years of estimated typical market liquidity (once an adjustment has been made to remove the historical DB buyer pool). But at the segment level, there is significant variation around this figure.

• Industrial holdings represent 2.5 years of liquidity in aggregate, 3.5 years for standard industrial and 1.5 years for distribution warehouses.
• The only other sector with more than a year’s worth of liquidity is residential, at 1.1 years.
• Retail holdings represent 0.9 years of supply—retail warehouses (1.7 years) and supermarkets (1.4 years) have the largest exposure at the segment level.
• Hotel and office exposure is minimal relative to typical liquidity.

Figure 4: Private DB holdings, value in £ in billions and typical years of investment supply

Figure 5: Private DB holdings, value in £ in billions and typical years of investment supply, by exposure category

Concentrations highlight potential risks and opportunities

When the trend of DB schemes derisking a few years ago came to the attention of real estate market participants, there was significant concen that a rapid, disorderly exit at scale by a traditional bullwark could lead to downard pressure on values. That situation did not materialize because the timescale for derisking turned out to be longer than anticipated, the size of the portfolio derisking was smaller than thought, and because the owners themselves are motivated to achieve the best outcome on exit. This context on the private DB real estate portfolio should be considered when analyzing the potential for future systemic risk.

The risks of a pricing shock caused by the sell down of the private DB real estate portfolio appears low. Exposure of £40.5 billion represents under a year’s typical supply, and all the signs so far are of an orderly withdrawal. For context, this is orders of magnitude lower than the volume of distressed debt in the U.K. banking system in the aftermath of the GFC. At that time, typical liquidity was lower than currently, while the slotting regime penalized banks from holding non-perfoming loans making them extremely eager to exit positions. Even then, there was no market-level hit to valuations, although Quantitative Easing provided a tailwind.

We are confident, therefore, that the U.K. will avoid widespread pricing pressure as private DB sells down. Even where there are pockets of concentration, we do not believe the risk is high. Investor appetite is strong for multi-let industrial and distribution warehouse, which represent the largest overweights relative to typical liquidity in the private DB real estate portfolio. Indeed by using the 2010-2019 period as the base level for setting typical liquidity, we may be understating demand for these segments of the market. The early part of that period, in particular, saw far less investment in the industrial sector generally than in 2020-2024.

As to opportunities, our analysis provides clear guidance on the scale and segment composition of the private DB real estate portfolio for investors looking to target assets from this group of motivated sellers. A significant proportion of the portfolio—though less than is the case for public DB—is in pooled and long-income strategies, which offer the potential for investment at scale via portfolio solutions and recapitalizations. Overall, the £20 billion held across industrial and residential, sectors, which investors are most strongly targeting overweight positions, represents an attractive amount of potential investable stock. There is also a decent pool of London offices and retail warehouses which are also currently attratctive to a range of buyers. The opportunity for acquiring shopping centers and leisure though is far smaller.

By combining data from a range of sources, we quantified the size of the DB real estate portfolio in total and concluded that the private DB part is likely to sell down over the next zero to fifteen years. At around £40 billion, the size of the portfolio is significant enough to provide opportunity to investors seeking entry into the U.K. but does not appear large enough to pose a systemic risk to pricing, especially given the orderly behavior of the sell-down process to date.