Should you regularly rebalance your portfolio?

I ask as I have revisited the WisdomTree Cloud Computing ETF (KLWD) held within my son’s SIPP.

Most indices are weighted towards market cap. But KLWD follows an equal-weighted index and is rebalanced every six months:

KLWD’s performance has trumped the large US indices:

That outperformance is almost certainly due to a focus on dynamic cloud-computing businesses.

But I wonder whether the regularly rebalancing of the ETF’s positions has helped.

KLWD was launched only in 2018, so the chart above involves a lot of theoretical ‘back testing’. Still, the post-launch performance is not too bad.

At the last count KLWD had 58 shares and each were re-weighted to 1.7% back in February.

One obvious issue with regular rebalancing is it stops you ‘running your winners’ – or as Peter Lynch told us to avoid: “cutting the flowers to water the weeds”.

That said, the approach may sensibly trim over-priced holdings to reinvest in out-of-favour positions. And it will prevent the ETF from becoming too dependent on just a handful of shares – a position the Nasdaq arguably faces.

In an effort to research further, I have looked back at my portfolio. Sadly I found no major evidence of the equal weighting significantly outperforming my actual results:

Annual return 2016 2017 2018 2019 2020 Compound
Actual (%) 7.6 10.5 (6.6) 13.1 16.9 46.8
Equal-weighted (%) 16.4 14.5 (4.4) 7.6 15.8 58.9

I assumed I had started 2016 with equal positions and then rebalanced at the end of every year. The associated sums are subject to numerous caveats, but the five-year compound outperformance is not enough to persuade me to look further and contemplate mechanical portfolio re-jigs.

So: should you regularly rebalance your portfolio? Probably not, but it may make no real difference if you do – unless of course you latch onto some really big winners. Which leaves me wondering how KLWD will perform over time. If the ETF really is harbouring a batch of ‘the next Microsofts’, my son will not enjoy the full multi-bagger exposure.

Maynard

When I have discussed this with wealth managers, or read about it in articles, it is always in the context of managing risk - so avoiding the possibility of one position growing so large that it overshadows the rest of the portfolio. Therefore if it collapses, the value of the portfolio is massively impaired. I have always found that rather implausible and over-defensive but maybe that is because I have never been in such a position. If I had bought Novacyt (NCYT) at 10p and held on till it reached £15 and now see it back at £4, perhaps I would have a different perspective! But I believe it is better to assess each position and work out whether the investment case still holds. And, from time to time, do some top-slicing when a share rockets too high too fast. The problem, as always, is information. With NCYT, a lot of people are foreseeing disaster but we have very little data until the accounts come out in June and they will be 6 months out of date.

The other factor is dealing costs. Did you include them in your analysis?

Agreed, as following mechanical rules is unlikely to have left NCYT untouched from 10p to 150p let alone to £15! A lot of investment success comes down to intuition and frequently any rules will tell you to do one thing while your assessment says to do another.

I suppose rebalancing may be effective if the expected returns from your shares are likely to be similar over time – that is perhaps the case with KLWD, as the constituents are all similar businesses. But there is every chance a handful of shares within KLWD will outperform the rest by miles, and so the fund won’t get the full exposure to their returns.

I should add that KLWD not only rebalances every six months, but reviews its index constituents every six months, too. Suitable cloud-computing businesses exhibiting slowing sales growth (less than 7%) are replaced by those exhibiting accelerating sales growth (15% or more). So a natural churn occurs, and therefore the fund is (in theory!) not consistently averaging down into ongoing under-performers. Maybe this churn, and not the rebalancing, has assisted the performance.

My analysis was very cursory and did not include dealing costs. Illiquidity would also be a problem rebalancing some of my shares on a regular basis. I think I will continue to depend on my own intuition/assessments, but check in on whether rebalancing is assisting KLWD – or whether its returns are correlated entirely to the success of cloud computing in general.

Maynard

I have looked further into this rebalancing matter. There could be an argument for regular rebalancing if your shares are likely to deliver similar returns over time. But rebalancing may not work as well if your shares are likely to deliver very different performances.

I have used SharePad to extract the annual share-price returns of every member of the FTSE 350 for the 10 years to 31 December 2020. 273 shares currently reside within the FTSE 350 with a 10-year record:

The following table lists the 10 best performers over the ten years, and their yearly returns:

Share 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
GAW (1) 5.3 49.4 3.8 (24.7) 12.5 22.0 269.0 15.5 101.0 83.5
AHT (2) 30.7 88.6 78.3 51.6 (2.9) 41.2 26.1 (17.8) 47.5 42.4
JD (3) (28.7) 9.8 113.0 39.3 105.0 52.7 5.8 3.8 140.0 2.7
AVON (4) 57.1 17.5 58.1 30.8 29.4 5.2 17.1 2.5 67.2 51.4
LIO (5) (14.5) 60.7 115.0 7.7 2.8 37.2 27.8 18.8 89.0 18.2
OCDO (6) (69.5) 59.3 410.0 (9.4) (24.0) (13.2) 50.4 98.9 61.9 78.8
LSEG (7) (5.1) 36.9 59.3 39.1 23.5 6.2 30.2 7.1 90.8 16.2
ENT (8) 15.9 88.3 52.2 35.3 (3.6) 38.6 44.0 (27.1) 31.2 28.2
FOUR (9) (14.3) 53.3 88.3 20.4 57.8 39.8 7.0 (2.9) 88.6 (26.3)
ATT (10) (9.3) 8.1 63.2 7.1 15.0 27.0 42.7 4.4 35.0 80.3

Investing a notional £100 in each would have resulted in a £15,076 portfolio at the end of 2020:

Untouched 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
GAW (1) 100 105 157 163 123 138 169 623 720 1,447 2,654
AHT (2) 100 131 247 440 666 647 914 1,152 947 1,397 1,990
JD (3) 100 71 78 167 232 476 727 769 798 1,915 1,966
AVON (4) 100 157 185 292 382 494 519 608 624 1,043 1,579
LIO (5) 100 86 137 295 318 327 449 573 681 1,287 1,522
OCDO (6) 100 31 49 248 224 171 148 223 443 717 1,282
LSEG (7) 100 95 130 207 288 355 377 491 526 1,004 1,167
ENT (8) 100 116 218 332 449 433 600 864 630 827 1,060
FOUR (9) 100 86 131 247 298 470 657 703 683 1,288 949
ATT (10) 100 91 98 160 171 197 250 357 372 503 907
Total 1,000 968 1,430 2,551 3,152 3,709 4,811 6,365 6,424 11,427 15,076

But had you rebalanced the portfolio at the end of every year, the total value would have been greater at £21,482:

Rebalanced 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
GAW (1) 100 105 145 148 219 392 516 1,961 933 1,792 2,866
AHT (2) 100 131 182 254 441 338 597 670 664 1,315 2,224
JD (3) 100 71 106 303 405 713 646 562 838 2,139 1,604
AVON (4) 100 157 114 225 380 450 445 622 829 1,490 2,365
LIO (5) 100 86 155 306 313 358 580 679 960 1,685 1,846
OCDO (6) 100 31 154 726 263 264 367 799 1,607 1,443 2,793
LSEG (7) 100 95 132 227 404 430 449 692 865 1,701 1,815
ENT (8) 100 116 182 217 393 335 586 765 589 1,169 2,002
FOUR (9) 100 86 148 268 350 549 591 569 785 1,681 1,151
ATT (10) 100 91 105 232 311 400 537 759 843 1,203 2,816
Total 1,000 968 1,424 2,907 3,480 4,230 5,316 8,080 8,914 15,619 21,482

This next table lists the 10 worst performers over the ten years, and their yearly returns:

Share 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
MTO (264) 3.8 8.2 20.9 (12.6) 11.9 (27.8) (13.9) (42.8) 31.1 (45.1)
RTN (265) 8.3 29.9 53.0 11.5 3.9 (52.7) (7.2) (35.1) 14.3 (60.6)
STAN (266) (18.3) 11.7 (13.6) (29.2) (38.5) 17.7 17.6 (21.9) 16.9 (34.6)
SRP (267) (14.7) 12.9 (6.7) (67.8) (27.4) 51.6 (31.0) (3.3) 69.4 (26.2)
FGP (286) (15.1) (38.1) (27.6) (13.7) 0.6 (3.5) 6.7 (24.5) 50.4 (41.0)
CNA (269) (12.8) 15.3 4.2 (19.8) (21.8) 7.3 (41.3) (1.8) (33.8) (47.8)
HMSO (270) (13.7) 35.6 2.8 20.5 (0.8) (4.5) (4.5) (39.8) (6.3) (82.4)
CPI (271) (9.8) 20.1 37.5 4.1 11.7 (56.0) (24.5) (54.2) 46.2 (76.1)
POG (272) (46.2) (41.3) (79.8) (79.7) 23.4 6.6 11.5 (18.5) 102.0 155.0
TLW (273) 11.2 (10.1) (32.2) 51.6 (60.0) 88.7 (22.4) (13.3) (64.3) (53.8)

Investing a notional £100 in each would have resulted in a £197 portfolio at the end of 2020:

Untouched 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
MTO (264) 100 104 112 136 119 133 96 83 47 62 34
RTN (265) 100 108 141 215 240 249 118 109 71 81 32
STAN (266) 100 82 91 79 56 34 40 48 37 43 28
SRP (267) 100 85 96 90 29 21 32 22 21 36 27
FGP (286) 100 85 53 38 33 33 32 34 26 39 23
CNA (269) 100 87 101 105 84 66 71 41 41 27 14
HMSO (270) 100 86 117 120 145 144 137 131 79 74 13
CPI (271) 100 90 108 149 155 173 76 58 26 39 9
POG (272) 100 54 32 6 1 2 2 2 2 3 8
TLW (273) 100 111 100 68 103 41 78 60 52 19 9
Total 1,000 893 951 1,006 965 896 681 588 402 422 197

But had you rebalanced the portfolio at the end of every year, the total value would have been greater at £401:

Rebalanced 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
MTO (264) 100 104 97 113 78 86 50 62 37 62 32
RTN (265) 100 108 116 143 100 80 33 67 41 54 23
STAN (266) 100 82 100 81 63 48 82 84 50 56 38
SRP (267) 100 85 101 87 29 56 106 49 62 81 43
FGP (286) 100 85 55 67 77 78 67 76 48 72 34
CNA (269) 100 87 103 97 72 60 75 42 63 31 30
HMSO (270) 100 86 121 96 108 77 67 68 38 45 10
CPI (271) 100 90 107 128 93 86 31 54 29 70 14
POG (272) 100 54 52 19 18 95 74 80 52 96 149
TLW (273) 100 111 80 63 135 31 132 56 55 17 27
Total 1,000 893 932 894 773 698 717 639 476 583 401

This next table lists the 10 middle performers over the ten years, and their yearly returns:

Share 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
DLN (133) (0.1) 35.0 18.5 21.0 21.7 (24.5) 12.5 (8.5) 40.6 (22.8)
ASL (134) (20.8) 38.8 57.4 (2.1) 11.3 (7.0) 19.6 (14.2) 35.3 (19.0)
SIG (135) (19.7) 25.3 43.8 12.3 (26.6) 49.8 23.4 (37.6) 16.3 22.0
COA (136) (17.4) 7.0 (1.6) (30.0) 17.9 120.0 63.6 (8.6) (8.5) (9.8)
LWDB (137) (6.5) 27.4 24.5 0.2 (6.0) 6.4 18.7 (14.1) 20.4 6.2
BRW (138) (13.7) 50.7 47.7 (2.3) 4.1 (1.6) 27.8 (17.3) 15.5 (18.1)
ADM (139) (43.8) 36.2 12.9 0.9 25.5 10.1 9.6 2.3 12.8 25.9
ABF (140) (6.3) 41.3 56.3 29.0 6.0 (17.9) 2.7 (27.6) 27.2 (12.9)
SAIN (141) (15.1) 8.2 13.6 (2.6) 4.8 23.9 13.6 (4.6) 21.4 8.9
EZJ (142) (10.7) 94.9 101.0 8.8 4.1 (42.2) 45.7 (24.5) 28.9 (41.7)

Investing a notional £100 in each would have resulted in a £1,933 portfolio at the end of 2020:

Untouched 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
DLN (133) 100 100 135 160 193 235 178 200 183 257 199
ASL (134) 100 79 110 173 169 189 175 210 180 243 197
SIG (135) 100 80 101 145 162 119 179 220 138 160 195
COA (136) 100 83 88 87 61 72 158 258 236 216 195
LWDB (137) 100 94 119 148 149 140 149 176 152 182 194
BRW (138) 100 86 130 192 188 195 192 246 203 235 192
ADM (139) 100 56 77 86 87 109 121 132 135 152 192
ABF (140) 100 94 132 207 267 283 232 239 173 220 191
SAIN (141) 100 85 92 104 102 106 132 150 143 173 189
EZJ (142) 100 89 174 350 381 396 229 334 252 325 189
Total 1,000 846 1,158 1,653 1,759 1,845 1,744 2,165 1,794 2,164 1,933

But had you rebalanced the portfolio at the end of every year, the total value would have been greater at £2,315:

Rebalanced 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
DLN (133) 100 100 114 137 192 200 132 219 221 287 190
ASL (134) 100 79 117 182 155 183 162 233 207 276 200
SIG (135) 100 80 106 166 178 121 261 241 151 237 301
COA (136) 100 83 91 114 111 194 384 319 220 187 223
LWDB (137) 100 94 108 144 159 154 186 231 207 245 262
BRW (138) 100 86 127 171 155 171 172 249 199 235 202
ADM (139) 100 56 115 130 160 206 192 214 247 230 311
ABF (140) 100 94 120 180 205 174 143 200 175 259 215
SAIN (141) 100 85 91 131 155 172 216 221 230 247 269
EZJ (142) 100 89 165 232 173 171 101 284 182 263 144
Total 1,000 846 1,155 1,586 1,642 1,745 1,949 2,412 2,039 2,467 2,315

So from this admittedly limited study, there seems to be a correlation between rebalancing and enhancing portfolio returns where the portfolio contains members delivering very similar long-term returns.

Mixing things up a little, this next table lists the top 5 and bottom 5 performers over the ten years, and their yearly returns:

Share 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
GAW (1) 5.3 49.4 3.8 (24.7) 12.5 22.0 269.0 15.5 101.0 83.5
AHT (2) 30.7 88.6 78.3 51.6 (2.9) 41.2 26.1 (17.8) 47.5 42.4
JD (3) (28.7) 9.8 113.0 39.3 105.0 52.7 5.8 3.8 140.0 2.7
AVON (4) 57.1 17.5 58.1 30.8 29.4 5.2 17.1 2.5 67.2 51.4
LIO (5) (14.5) 60.7 115.0 7.7 2.8 37.2 27.8 18.8 89.0 18.2
CNA (269) (12.8) 15.3 4.2 (19.8) (21.8) 7.3 (41.3) (1.8) (33.8) (47.8)
HMSO (270) (13.7) 35.6 2.8 20.5 (0.8) (4.5) (4.5) (39.8) (6.3) (82.4)
CPI (271) (9.8) 20.1 37.5 4.1 11.7 (56.0) (24.5) (54.2) 46.2 (76.1)
POG (272) (46.2) (41.3) (79.8) (79.7) 23.4 6.6 11.5 (18.5) 102.0 155.0
TLW (273) 11.2 (10.1) (32.2) 51.6 (60.0) 88.7 (22.4) (13.3) (64.3) (53.8)

Investing a notional £100 in each would have resulted in a £9,764 portfolio at the end of 2020:

Untouched 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
GAW (1) 100 105 157 163 123 138 169 623 720 1,447 2,654
AHT (2) 100 131 247 440 666 647 914 1,152 947 1,397 1,990
JD (3) 100 71 78 167 232 476 727 769 798 1,915 1,966
AVON (4) 100 157 185 292 382 494 519 608 624 1,043 1,579
LIO (5) 100 86 137 295 318 327 449 573 681 1,287 1,522
CNA (269) 100 87 101 105 84 66 71 41 41 27 14
HMSO (270) 100 86 117 120 145 144 137 131 79 74 13
CPI (271) 100 90 108 149 155 173 76 58 26 39 9
POG (272) 100 54 32 6 1 2 2 2 2 3 8
TLW (273) 100 111 100 68 103 41 78 60 52 19 9
Total 1,000 979 1,262 1,805 2,210 2,508 3,141 4,018 3,969 7,250 9,764

But had you rebalanced the portfolio at the end of every year, the total value would have been lower at £4,168:

Rebalanced 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
GAW (1) 100 105 146 127 119 193 230 835 330 515 700
AHT (2) 100 131 185 217 240 167 266 285 235 378 543
JD (3) 100 71 107 260 221 352 288 239 297 615 392
AVON (4) 100 157 115 193 207 222 198 265 293 428 577
LIO (5) 100 86 157 262 171 176 259 289 340 484 451
CNA (269) 100 87 113 127 127 134 202 133 281 170 199
HMSO (270) 100 86 133 125 191 170 180 216 172 240 67
CPI (271) 100 90 118 168 165 192 83 171 131 375 91
POG (272) 100 54 57 25 32 212 201 252 233 517 972
TLW (273) 100 111 88 83 240 69 356 176 248 91 176
Total 1,000 979 1,219 1,586 1,715 1,885 2,263 2,861 2,562 3,813 4,168

And this table lists a spread of 10 shares over the ten years, and their yearly returns:

Share 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
GAW (1) 5.3 49.4 3.8 (24.7) 12.5 22.0 269.0 15.5 101.0 83.5
HPVE (31) 7.9 20.2 30.3 28.7 14.4 26.6 7.7 8.4 33.4 8.2
III (61) (44.9) 20.0 77.3 16.9 7.0 46.2 29.8 (15.3) 42.0 5.5
JESC (92) (25.6) 12.1 57.0 (10.6) 42.3 0.7 43.3 (21.9) 16.5 25.9
WTAN (122) (12.9) 11.8 33.0 12.6 3.5 15.6 19.6 (10.0) 19.2 (0.4)
ROR (152) 5.6 31.9 12.8 (19.0) (21.5) 32.0 10.7 (7.2) 35.3 (5.1)
LRE (182) 31.0 7.0 4.6 (30.9) 12.1 10.5 (1.7) (11.3) 26.8 (5.7)
PTEC (213) (33.5) 51.3 72.8 (6.5) 20.8 (0.8) 4.2 (55.3) 3.1 1.1
SBRY (243) (19.5) 13.9 5.8 (32.4) 4.9 (3.7) (3.2) 9.8 (13.2) (1.9)
TLW (273) 11.2 (10.1) (32.2) 51.6 (60.0) 88.7 (22.4) (13.3) (64.3) (53.8)

Investing a notional £100 in each would have resulted in a £4,504 portfolio at the end of 2020:

Untouched 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
GAW (1) 100 105 157 163 123 138 169 623 720 1,447 2,654
HPVE (31) 100 108 130 169 217 249 315 339 368 490 531
III (61) 100 55 66 117 137 147 214 278 236 335 353
JESC (92) 100 74 83 131 117 167 168 240 188 219 275
WTAN (122) 100 87 97 130 146 151 175 209 188 224 223
ROR (152) 100 106 139 157 127 100 132 146 135 183 174
LRE (182) 100 131 140 147 101 114 126 123 109 139 131
PTEC (213) 100 67 101 174 163 196 195 203 91 94 95
SBRY (243) 100 81 92 97 66 69 66 64 70 61 60
TLW (273) 100 111 100 68 103 41 78 60 52 19 9
Total 1,000 925 1,106 1,352 1,300 1,371 1,636 2,286 2,157 3,209 4,504

But had you rebalanced the portfolio at the end of every year, the total value would have been lower at £2,764:

Rebalanced 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
GAW (1) 100 105 138 116 106 157 176 659 280 438 480
HPVE (31) 100 108 111 145 182 159 183 192 263 291 283
III (61) 100 55 111 198 165 149 211 232 205 309 276
JESC (92) 100 74 104 175 126 198 145 256 189 254 329
WTAN (122) 100 87 103 148 159 144 167 214 218 260 260
ROR (152) 100 106 122 126 114 109 190 198 225 295 248
LRE (182) 100 131 99 117 98 156 159 175 215 276 247
PTEC (213) 100 67 140 193 132 168 143 186 108 225 264
SBRY (243) 100 81 105 118 95 146 139 173 266 189 256
TLW (273) 100 111 83 76 214 56 272 139 210 78 121
Total 1,000 925 1,116 1,413 1,392 1,442 1,785 2,423 2,179 2,614 2,764

These more mixed portfolios suggest a correlation between rebalancing and reducing portfolio returns where the portfolio contains members delivering very different long-term returns.

What could this mean for KLWD? I suspect the fund’s great performance to date has been due to the majority of these cloud-computing stocks doing well and the rebalancing perhaps trimming some temporary outperformers to reinvest in some temporary underperformers.

The real test will come over time, as I am sure the sector will split into huge winners, so-so middlers and complete losers… at which point the rebalancing may compromise the returns of the huge winners. We shall see.

Maynard

Fascinating research Maynard - many thanks.

Have brokerage / stamp duty / price spreads been taken into account in your rebalancing figures?

Your research clearly indicates to me that unless an investor has a crystal ball regarding future returns of individual lines and is correct in assuming his portfolio consists solely of members with similar future return prospects (unlikely) then rebalancing is likely to have a negative effect.

It has certainly convinced me to leave regular rebalancing well alone as I simply don’t have a crystal ball.

Snazzy

That certainly gives food for thought. I imagine that dealing costs at something like 2% per annum would have a sizeable impact on the differential performance of the rebalanced portfolios. Maybe it would be better just to trim your under- performers but, in your first portfolio, that would have meant trimming GAW in some years and therefore losing some of its alpha. It certainly suggests that following rules might not be the best approach.

Closer inspection, I see that GAW’s drop in 2014 led to a big increase in position, which fed through in later years. In a portfolio where not every share is shooting the lights out, topping up your under performers might not be so beneficial

Hi Snazzy,

The stats do not include costs, spreads and dividends.

Yes, I would agree.

A few more thoughts:

The 10 top winners were not all ‘quality’ companies 10 years ago. Certainly GAW and LIO had experienced troubles in the few years beforehand. Anyone owning those 10 shares at the end of 2010 would have reasonably assumed a spread of performances would occur during the next decade, and therefore would have thought to ‘run their winners’ in order to counterbalance the also-rans. I doubt anyone would have imagined they would in fact all 10-bag or more.

With the 10 top losers, the rebalancing effect only really kicked in at the start of 2019 when POG took off. Otherwise it was touch and go between the strategies for the first eight years. All it took was one big recovery (albeit after a crushing plunge!) to make the 100% difference between the end results (£197 vs £401). Mind you being down 60% instead of 80% may be a moot point.

With the 10 middle performers, rebalancing only started to outperform from 2016 – not ideal. But I note 8 of the 10 holdings did better being rebalanced than being left untouched.

With the spread of 10 performers, rebalancing actually outperformed slightly until 2019, at which point GAW took off and made all the difference to the end result. This perhaps is the reality most investors experience – a spread of performers, and holding tightly to one particular share that does extremely well to outperform.

Maynard

Hi Diogenes,

In that first portfolio a major difference was surprisingly made with ATT. Untouched, ATT turned £100 into £907, but rebalanced turned £100 into £2,816 – to become almost the largest holding.

I think this performance was due to ATT reporting a loss in the first year, then going on to deliver nine straight gains – with the latter years particularly strong.

In fact, nine of the 10 shares in the top winners table ended with a higher valuation as a result of rebalancing. I guess that comes from ensuring you have 10 big winners – every year you are guaranteed a large gain from one or two to spread over to the others to compound in future years!

Maynard