Understanding Allocation Benefit: The U.K., sector or stock?

Executive Summary and Key Calls

Key call 1: Typically, the difference between being a good stock-picker or good allocator is small…

We calculated the difference in performance that might be attributed to exceptional top-down allocation or bottom-up stock-picking and found that over five-year periods, the delta is usually small. From 1990-2017, there were 25 instances out of a possible 28 where the difference between the two was less than +/-2% and 18 where the difference was less than +/-1%.

Key call 2: ...but top-down allocation drove recent performance by an unprecedented degree

With the significant recent divergence in sector performance, allocation became more important. In the seven years, 2017-2023, allocation benefit exceeded selection benefit by a cumulative 36.4%, or an average 5.2% (compared to its long-run average of -1.3%).

Key call 3: Stock-picking is due for a resurgence

We expect selection to drive performance to a greater extent in the coming years. The difference in investment returns between median and upper quartile assets will be exacerbated by a bifurcation in demand affecting all sectors, while discounted entry prices from distressed or motivated sellers will be an additional source of outperformance for well-connected investors. 

A way of quantifying the drivers of performance

Top-down vs. bottom-up? Or….researchers vs. fund managers?

A mainstay of MSCI portfolio performance measurement has long been attribution analysis, which seeks to quantify the extent to which a fund’s performance relative to its benchmark was due to allocation—being underweight to underperforming sectors and/or overweight to outperforming ones—or stock selection—investing in better-performing assets within sectors. The big question is whether relative performance is a function of top-down or bottom-up decisions.

In reality, a blend of the two usually exists, and the best investment and research processes are collaborative across departments rather than siloed. There are, however, periods when the extent to which being good at one or other discipline offers a greater or lesser opportunity to outperform; when returns can be enhanced more by being better at stock-picking than sector allocation or vice versa.

Quantifying allocation and selection benefit

To understand the extent to which stock-picking and sector allocation played, or could have played, a part in determining investment performance, we need a way to quantify the potential impact of each. Figure 1 sets out our approach, using 2024 as an example.

• For sector allocation, or allocation benefit as we have termed it, we take the sector average total return and calculate the average uplift an investor would get by being in the top-performing sector rather than any other. In 2024, this was 3.4%, which was the average of the difference between industrial returns of 7.9% and retail (7.8%), office (1.0%), residential (2.8%) and hotel (3.2%).
• For stock-picking, or selection benefit as we have termed it, we take the average difference between the total return on the median asset and the 75th percentile asset within each sector. In 2024, this was 4.0% (and ranged from 3.2%-5.3% depending on the sector).

Performing this calculation over time, and for longer rolling periods (e.g., five years), reveals when and to what extent allocation or selection benefit has been the more dominant driver of portfolio performance.

Figure 1: Calculating allocation and stock selection impact, 2024 example*

2017-2023 was all about allocation, to an almost unprecedented degree

Allocation was king, but was dethroned in 2024

In Figure 2 we show allocation and selection benefit since 1981 (the full extent of the MSCI Annual Digest). Figure 2a displays a rolling one-year view and Figure 2b a five-year view. On a one-year basis, our analysis shows that:

• Over the full history, on a one year basis, the average allocation benefit has been 6.2%, a little lower than the average selection benefit of 7.5%.
• In roughly two thirds of the years (30), selection benefit has exceed allocation benefit (14).
• However, allocation benefit has recently been dominant. In the seven years, 2017-2023, allocation benefit exceeded selection benefit by a cumulative 36.4%, or an average 5.2% (compared to its long-run average of -1.3%).
• This is almost unprecedented; the last time one factor was so consistently dominant was in the period 1981-1990, when selection benefit exceeded allocation benefit for ten consecutive years by a cumulative 50.0%....
• 2024 saw the end of the run however; selection benefit exceeded allocation benefit by 0.6%.

Figure 2a: Allocation and stock selection impact on one-year total return*

Figure 2b: Allocation and stock selection impact on five-year total return*

Switching to a five-year view, we of course observe many of the same trends. Although different sectors may outperform in individual years through a five year period, allocation benefit is not simply the sum of one year numbers. At the asset level, underperformers in one year may be outperformers the next (or underperformers by a lesser extent). Selection benefit is similarly subtly different.

• Over five-year periods, similar overall patterns are in evidence: allocation benefit has dominated of late, whereas selection benefit had its moment in the sun in the 1980s.
• In the intervening years, 1990-2017, the proximity of the two benefits rather than their divergence is interesting. There were 25 instances out of a possible 28 where the difference between the two figures was less than +/-2% and 18 instances where the difference was less than +/-1%.
• The change in level pattern is also interesting. It is perhaps more observable in the selection benefit series, which sees a definite step down in prominence around the early 2000s—originally trending close to 5% moving to a trend of close to 3%. Greater market transparency and the evolution of the MSCI sample likely both play a part in this occurrence.

Is selection now set to dominate for a period?

The difference between allocation benefit and selection benefit is shown in Figure 3—again, the charts are done on a one- and five-year basis. These charts show more clearly the pattern in evidence over the past 40+ years—the recent supremacy of allocation benefit, the long-ago dominance of selection benefit and the intervening majority period of one factor being a little more meaningful for a few years, before the other factor switches places.

Figure 3a: Allocation and stock selection impact on one-year total return*

Figure 3b: Allocation and stock selection impact on five-year total return*

Assuming, as the direction of travel on both the one and five year charts demonstrate, that the recent era of overwhelming dominance by the allocation benefit has ended, what does the future hold? Since we are ruling out one possibility, which of the other two do we believe is more likely to emerge?

The case for selection benefit being substantially greater than allocation benefit for a prolonged period can arguably be made in one word—bifurcation—specifically the bifurcation of tenant demand for the best assets vs. the rest. In the last few years, we saw prime office rent growth in double-digits in comparison with secondary stock where vacancy soared. Occupiers are more discerning on location, asset quality and functionality, which has led to a wide divergence in performance between modern and legacy offices. This bifurcation is no longer unique to offices. Net absorption in industrial space, for example, is massively skewed toward modern stock. In retail the playing out of winners and losers in a world of ever-increasing online penetration is far from finished. We believe that in the near term, the difference in investment returns between median and upper quartile assets will grow.

The case for the middle ground, of a return to the long period when there wasn’t much difference (certainly on a five-year rolling view) between allocation and selection benefit, is perhaps also summed up in a single word—chartism. The periods of dominance by allocation benefit in the late 2010s and selection benefit in the 1980s are outliers; far more common is the 1990-2015 run of minimal difference between the two factors. On a purely historical chartist basis, it can be argued that it is more likely that this scenario will recur and that portfolio management will return to the long era of being a roughly equal balance between top-down and bottom-up calls.

Benchmarking allocators and stock-pickers

How good have each needed to be through time?

The above analysis has allowed us to think about what has had the greater impact on investment performance over the last 45 years—allocation or selection. A further area to explore is what level of selection ability equates to a given level of allocation ability.

We have, somewhat arbitrarily perhaps, equated an allocator who can pick the best sector with a stock selector who can identify 75th percentile assets. But is this fair? Are these roughly the same? Figure 4 suggests that they are. As we noted earlier in this piece, the allocation benefit has averaged 6.2% over time, while the selection benefit (set at the 75th percentile) has been 7.5%. Figure 4 shows what the selection benefit would be had we chosen all other percentiles available to us (note: MSCI only releases data in increments of five percentiles). The benefit from moving from median to 70th percentile is 5.5% on average, implying that a benefit of 6.2% (to match exactly the allocation benefit) might be at the 72nd percentile. Since the numbers are close, we retain the 75th percentile as broadly representing the level of asset selection that is needed to match the allocation benefit of identifying the top performing sector.

Figure 4: Selection benefit depending on percentile to which selection benefit moves from median*

A final way to compare the two disciplines is shown in Figure 5, which for each one- and five-year period finds the percentile at which the selection benefit matches the allocation benefit. This varies considerably through time, in line with the swings in importance of allocation and selection discussed earlier and helps to contextualize how much better at selection an investor had to have to match allocation benefit when allocation was in the ascendancy or vice versa. To match the five-year allocation benefit in the periods ending 2020-2024 via selection, investors would have needed to identify 95th percentile assets, rather than the 70th-80th percentile more typical from the 1990s to 2010s, or the 55th-65th percentile seen in the 1980s.

Figure 5: Equalizing allocation benefit and selection benefit by changing percentile to which selection benefit moves from median*

In summary…

Back to basics?

The last few years saw portfolio performance determined by sector allocation to an unprecedented degree—but this looks set to change. Successful real estate investment has almost always been a balance of allocation and stock selection, and the importance of this latter discipline in driving returns is already re-emerging as the new cycle takes shape. Understanding the varying potential for income growth of existing assets and potential acquisitions, and managing exposure accordingly, will once again be a cornerstone of investors’ success. And investors who can drive additional performance at the asset level by managing to enter at a favorable, discounted, price by taking advantage of distressed or motivated sales, for example, could also be at an advantage.