WEBVTT

1
00:00:00.020 --> 00:00:00.180
Hello,

2
00:00:00.280 --> 00:00:02.281
welcome back to Papers with Backtest podcast.

3
00:00:02.623 --> 00:00:02.863
Today,

4
00:00:02.982 --> 00:00:05.645
we dive into another Algo trading research paper.

5
00:00:05.785 --> 00:00:09.188
And today we're looking at adaptive asset allocation,

6
00:00:09.391 --> 00:00:09.930
a primer.

7
00:00:10.328 --> 00:00:11.430
It's by Adam Butler,

8
00:00:11.508 --> 00:00:12.234
Michael Philbrick,

9
00:00:12.312 --> 00:00:14.836
and Rodrigo Gordillo from Resolve Asset Management.

10
00:00:14.938 --> 00:00:15.078
Right,

11
00:00:15.094 --> 00:00:17.062
the 2015 revision specific.

12
00:00:17.078 --> 00:00:17.219
Well,

13
00:00:17.234 --> 00:00:17.891
that's the one.

14
00:00:18.000 --> 00:00:21.500
And it really dives into a core problem with,

15
00:00:21.562 --> 00:00:21.875
let's say,

16
00:00:21.969 --> 00:00:25.625
traditional investing ideas like modern portfolio theory,

17
00:00:26.141 --> 00:00:26.469
MPT.

18
00:00:27.008 --> 00:00:27.128
OK.

19
00:00:27.388 --> 00:00:27.669
MPT.

20
00:00:27.829 --> 00:00:28.449
What's the issue,

21
00:00:28.550 --> 00:00:28.910
Grace?

22
00:00:29.010 --> 00:00:29.131
Well,

23
00:00:29.132 --> 00:00:32.431
the paper argues that MPT often leans really heavily on these,

24
00:00:32.615 --> 00:00:32.853
you know,

25
00:00:33.096 --> 00:00:36.416
super long term average figures for returns,

26
00:00:36.455 --> 00:00:36.978
for risk.

27
00:00:36.979 --> 00:00:38.463
They're predicting decades out.

28
00:00:38.517 --> 00:00:39.103
Exactly.

29
00:00:39.283 --> 00:00:40.424
And the authors kind of ask,

30
00:00:40.502 --> 00:00:44.361
are those long term guesses really the best foundation for building a portfolio now?

31
00:00:44.486 --> 00:00:44.627
Uh-huh.

32
00:00:44.861 --> 00:00:46.033
And they bring up that acronym,

33
00:00:46.064 --> 00:00:46.267
right?

34
00:00:46.330 --> 00:00:46.783
GIGO.

35
00:00:47.096 --> 00:00:47.205
Yeah.

36
00:00:47.314 --> 00:00:47.892
Garbage in,

37
00:00:47.955 --> 00:00:48.455
garbage out.

38
00:00:48.502 --> 00:00:49.236
It's a classic.

39
00:00:49.533 --> 00:00:50.017
Basically,

40
00:00:50.158 --> 00:00:51.736
if the data you feed your model

41
00:00:51.948 --> 00:00:53.931
those long term return forecasts,

42
00:00:54.009 --> 00:00:54.931
volatility estimates,

43
00:00:54.970 --> 00:00:55.572
correlations.

44
00:00:55.591 --> 00:00:58.013
If that data is garbage or maybe just,

45
00:00:58.173 --> 00:00:58.415
you know,

46
00:00:58.693 --> 00:01:00.497
not very accurate for the current environment.

47
00:01:00.818 --> 00:01:02.880
Then the portfolio that comes out the other end,

48
00:01:04.021 --> 00:01:04.279
well,

49
00:01:04.341 --> 00:01:05.583
it might not be optimal.

50
00:01:06.021 --> 00:01:08.669
It could even be counterproductive for you as an investor.

51
00:01:08.904 --> 00:01:09.341
Makes sense.

52
00:01:09.607 --> 00:01:11.685
If you build based on flawed assumptions.

53
00:01:11.794 --> 00:01:14.044
Your results probably won't match your expectations.

54
00:01:14.045 --> 00:01:17.060
And they show this with figure two in the paper.

55
00:01:17.376 --> 00:01:20.358
Looking at stock versus bond excess returns.

56
00:01:20.419 --> 00:01:20.559
Right.

57
00:01:20.579 --> 00:01:21.401
That chart was interesting.

58
00:01:21.402 --> 00:01:23.401
It wasn't just a straight line up for stocks.

59
00:01:23.643 --> 00:01:24.104
Not at all.

60
00:01:24.182 --> 00:01:26.503
It shows how much that relationship fluctuates.

61
00:01:26.542 --> 00:01:30.588
There were actually pretty long periods where bonds did better than stocks,

62
00:01:30.971 --> 00:01:35.526
which kind of flies in the face of the simple stocks always outperform long term idea.

63
00:01:35.698 --> 00:01:41.917
So those long term averages we hear about aren't always reliable predictors for shorter or even medium time frames.

64
00:01:42.120 --> 00:01:42.620
Precisely.

65
00:01:42.745 --> 00:01:44.667
And then there's figure three looking at like.

66
00:01:44.820 --> 00:01:46.960
actual investor holding periods.

67
00:01:47.341 --> 00:01:49.181
People don't hold things for 30 years usually.

68
00:01:49.579 --> 00:01:50.060
Often not,

69
00:01:50.122 --> 00:01:50.240
no,

70
00:01:50.400 --> 00:01:51.341
maybe just a few years,

71
00:01:51.603 --> 00:01:52.060
typically,

72
00:01:52.583 --> 00:01:56.806
which again creates a mismatch if your whole strategy is built on multi-decade averages.

73
00:01:57.103 --> 00:01:57.220
So,

74
00:01:57.384 --> 00:01:57.603
okay,

75
00:01:57.665 --> 00:01:59.704
this sets the stage for adaptive allocation.

76
00:02:00.040 --> 00:02:00.626
Exactly.

77
00:02:00.736 --> 00:02:02.720
Instead of static long-term guesses,

78
00:02:02.798 --> 00:02:06.111
the paper explores using shorter-term observed market data,

79
00:02:06.657 --> 00:02:07.767
strategies that react.

80
00:02:08.032 --> 00:02:08.986
And for this deep dive,

81
00:02:09.001 --> 00:02:11.892
we're really going to focus on the specific trading rules they tested,

82
00:02:12.470 --> 00:02:12.767
and,

83
00:02:12.923 --> 00:02:13.329
crucially,

84
00:02:13.816 --> 00:02:15.117
The backtest results they found.

85
00:02:15.358 --> 00:02:15.478
Yep.

86
00:02:15.558 --> 00:02:17.100
Let's get into the mechanics and the numbers.

87
00:02:17.139 --> 00:02:17.340
Okay.

88
00:02:17.420 --> 00:02:17.881
First up,

89
00:02:19.483 --> 00:02:20.162
the baseline.

90
00:02:20.561 --> 00:02:21.904
The equal weight portfolio.

91
00:02:22.264 --> 00:02:22.983
How did that work?

92
00:02:23.248 --> 00:02:23.889
Pretty simple stuff.

93
00:02:23.904 --> 00:02:24.209
They took

94
00:02:24.647 --> 00:02:26.506
10 global asset classes.

95
00:02:26.670 --> 00:02:27.811
Like stocks from the US,

96
00:02:27.967 --> 00:02:28.248
Europe,

97
00:02:28.311 --> 00:02:28.811
Japan,

98
00:02:29.108 --> 00:02:29.850
emerging markets.

99
00:02:30.115 --> 00:02:30.334
Right.

100
00:02:30.428 --> 00:02:31.100
Plus REITs,

101
00:02:31.272 --> 00:02:32.694
both US and international,

102
00:02:32.975 --> 00:02:34.053
different US treasuries,

103
00:02:34.100 --> 00:02:35.334
like intermediate and long-term,

104
00:02:35.381 --> 00:02:36.522
plus commodities and gold.

105
00:02:36.740 --> 00:02:37.537
A broad mix.

106
00:02:37.897 --> 00:02:39.100
And just split the money equally.

107
00:02:39.256 --> 00:02:39.694
10%

108
00:02:39.740 --> 00:02:40.100
in each,

109
00:02:40.350 --> 00:02:41.225
rebalanced every month.

110
00:02:41.600 --> 00:02:41.959
That's it.

111
00:02:42.162 --> 00:02:43.084
The sort of naive...

112
00:02:43.220 --> 00:02:43.600
approach,

113
00:02:43.741 --> 00:02:44.361
as they call it.

114
00:02:44.542 --> 00:02:45.343
And the back test.

115
00:02:45.462 --> 00:02:47.946
This is from 1995 to 2014.

116
00:02:48.044 --> 00:02:48.927
Over that period,

117
00:02:48.946 --> 00:02:50.887
it gave 8.1%

118
00:02:50.888 --> 00:02:52.052
compound annual return.

119
00:02:52.231 --> 00:02:54.247
Volatility was 11.2%.

120
00:02:54.248 --> 00:02:54.348
Okay,

121
00:02:54.387 --> 00:02:54.833
sharp ratio.

122
00:02:54.834 --> 00:02:55.731
0.72.

123
00:02:56.169 --> 00:02:57.450
And the maximum drawdown,

124
00:02:57.473 --> 00:02:58.872
the biggest peak to trough loss,

125
00:02:58.934 --> 00:03:00.075
was pretty significant,

126
00:03:00.419 --> 00:03:01.825
negative 39.2%.

127
00:03:01.841 --> 00:03:02.059
Ouch.

128
00:03:02.497 --> 00:03:03.341
Almost 40%

129
00:03:03.387 --> 00:03:04.184
down at its worst.

130
00:03:04.325 --> 00:03:05.294
So that's our starting point.

131
00:03:05.544 --> 00:03:06.559
That's the benchmark.

132
00:03:06.966 --> 00:03:07.091
Now,

133
00:03:07.092 --> 00:03:10.012
the first adaptive strategy they tested was volatility weighted.

134
00:03:10.622 --> 00:03:10.887
All right.

135
00:03:10.888 --> 00:03:11.341
How does that

136
00:03:11.700 --> 00:03:12.747
Instead of equal weight,

137
00:03:13.048 --> 00:03:18.433
The allocation changes monthly based on each asset's volatility over the previous 60 days.

138
00:03:18.574 --> 00:03:19.914
So inverse volatility.

139
00:03:20.414 --> 00:03:22.215
More volatile assets get less weight.

140
00:03:22.300 --> 00:03:22.777
Exactly.

141
00:03:23.058 --> 00:03:29.629
The idea is to make each asset contribute roughly the same amount of risk or volatility to the total portfolio.

142
00:03:30.160 --> 00:03:30.691
With a cap,

143
00:03:30.769 --> 00:03:31.004
though,

144
00:03:31.410 --> 00:03:31.722
max

145
00:03:32.066 --> 00:03:33.363
100% exposure.

146
00:03:33.863 --> 00:03:34.254
Interesting.

147
00:03:34.269 --> 00:03:36.191
So managing risk more actively,

148
00:03:36.285 --> 00:03:37.144
how did that perform?

149
00:03:37.410 --> 00:03:40.675
The compound return actually ticked up a bit to 8.5%.

150
00:03:40.788 --> 00:03:42.490
But the real story was the volatility.

151
00:03:42.690 --> 00:03:42.950
Lower.

152
00:03:43.151 --> 00:03:43.311
Yeah,

153
00:03:43.510 --> 00:03:44.432
significantly lower,

154
00:03:44.573 --> 00:03:46.272
down to 8.6%.

155
00:03:46.315 --> 00:03:46.456
Ah,

156
00:03:46.612 --> 00:03:48.776
so that boosts the risk-adjusted return.

157
00:03:48.838 --> 00:03:49.034
Right.

158
00:03:49.198 --> 00:03:52.299
The Sharpe ratio improved quite a bit to 0.99.

159
00:03:52.643 --> 00:03:55.745
And maybe most importantly for someone watching their account balance.

160
00:03:55.924 --> 00:03:56.643
The drawdown.

161
00:03:56.831 --> 00:03:57.268
Much better.

162
00:03:57.424 --> 00:04:01.862
Maximum drawdown was cut substantially down to Tenegr 24.2%.

163
00:04:01.863 --> 00:04:02.034
Okay,

164
00:04:02.049 --> 00:04:04.315
that's a big difference from nearly negative 40%.

165
00:04:04.316 --> 00:04:07.752
Shows the power of just managing risk based on recent volatility.

166
00:04:07.753 --> 00:04:07.877
See,

167
00:04:07.893 --> 00:04:08.237
definitely.

168
00:04:08.315 --> 00:04:09.737
It smooths the ride quite a bit.

169
00:04:10.081 --> 00:04:10.377
Next.

170
00:04:10.548 --> 00:04:11.409
they brought in momentum.

171
00:04:11.929 --> 00:04:15.133
Strategy three was top five equal weight by six month momentum.

172
00:04:15.333 --> 00:04:15.473
Yep.

173
00:04:15.993 --> 00:04:20.880
So here every month you rank all 10 assets based on their total return over the past six months.

174
00:04:21.137 --> 00:04:22.442
The trend following idea.

175
00:04:22.637 --> 00:04:23.481
What's been going up?

176
00:04:23.684 --> 00:04:24.200
Kind of.

177
00:04:24.278 --> 00:04:26.544
You pick the top five performers from that list.

178
00:04:26.700 --> 00:04:27.778
And just hold those five.

179
00:04:27.825 --> 00:04:29.153
Hold those five in equal weight.

180
00:04:29.154 --> 00:04:29.481
So 20%

181
00:04:29.887 --> 00:04:30.731
each for the next month.

182
00:04:31.106 --> 00:04:33.372
Then you re-rank and rebalance again.

183
00:04:33.544 --> 00:04:34.825
Focusing on the recent winners.

184
00:04:35.091 --> 00:04:36.137
How did that strategy do?

185
00:04:36.247 --> 00:04:38.294
This is where the returns really started to pick up.

186
00:04:38.403 --> 00:04:39.653
Compound return jumped to

187
00:04:39.700 --> 00:04:41.362
13.0%.

188
00:04:41.482 --> 00:04:41.701
Wow,

189
00:04:41.903 --> 00:04:42.684
quite a leap from the

190
00:04:43.024 --> 00:04:43.883
8ish percent range.

191
00:04:44.203 --> 00:04:45.125
What about volatility?

192
00:04:45.485 --> 00:04:48.625
Volatility was actually similar to the basic equal weight,

193
00:04:48.672 --> 00:04:50.071
around 11.0%.

194
00:04:50.172 --> 00:04:50.391
Okay,

195
00:04:50.453 --> 00:04:52.289
so higher returns for similar volator.

196
00:04:52.914 --> 00:04:53.813
That means a better sharp.

197
00:04:54.297 --> 00:04:54.797
Much better.

198
00:04:55.157 --> 00:04:57.024
Sharp ratio hit 1.17.

199
00:04:57.497 --> 00:04:59.657
And the max drawdown also improved further,

200
00:04:59.877 --> 00:05:02.239
down to Negness 21.7%.

201
00:05:02.477 --> 00:05:02.899
Interesting.

202
00:05:02.938 --> 00:05:08.321
So focusing on momentum gave more return and didn't really add much risk compared to the baseline,

203
00:05:08.641 --> 00:05:10.516
even slightly reducing the worst drawdown.

204
00:05:10.704 --> 00:05:11.063
Seems so.

205
00:05:11.563 --> 00:05:14.977
It suggests there's value in following recent trends,

206
00:05:15.118 --> 00:05:16.274
at least in this context.

207
00:05:16.805 --> 00:05:16.961
Now,

208
00:05:17.102 --> 00:05:20.758
the next logical step seems to be combining the previous two ideas,

209
00:05:21.227 --> 00:05:22.649
momentum and volatility weighting.

210
00:05:22.821 --> 00:05:23.305
Exactly.

211
00:05:23.727 --> 00:05:24.383
Strategy four,

212
00:05:25.118 --> 00:05:25.899
top five.

213
00:05:26.045 --> 00:05:28.849
by six-month momentum volatility weighted.

214
00:05:29.048 --> 00:05:30.388
So same first step,

215
00:05:30.708 --> 00:05:33.110
pick the top five based on six-month returns.

216
00:05:33.212 --> 00:05:33.431
Right.

217
00:05:33.872 --> 00:05:35.575
But then instead of weighting them equally,

218
00:05:36.075 --> 00:05:40.622
you apply that inverse volatility logic we talked about earlier to just those five assets.

219
00:05:40.763 --> 00:05:40.942
Ah,

220
00:05:41.278 --> 00:05:42.474
so you pick the current leaders,

221
00:05:42.755 --> 00:05:46.989
but then you scale their position size based on how bumpy their recent ride has been.

222
00:05:47.099 --> 00:05:47.521
Precisely.

223
00:05:47.880 --> 00:05:51.364
Aiming for equal volatility contribution among the chosen top five.

224
00:05:51.771 --> 00:05:52.849
It's a neat combination,

225
00:05:52.958 --> 00:05:54.380
trying to get the best of both worlds,

226
00:05:54.427 --> 00:05:55.271
capture momentum,

227
00:05:55.485 --> 00:05:57.127
but manage the associated risk.

228
00:05:57.267 --> 00:05:57.527
Okay,

229
00:05:57.566 --> 00:05:59.349
how did that combo work out in the backtest?

230
00:05:59.531 --> 00:05:59.949
Even better.

231
00:06:00.269 --> 00:06:03.273
Compound return nudged up again to 14.0%,

232
00:06:03.750 --> 00:06:05.718
and the volatility actually decreased this time,

233
00:06:05.953 --> 00:06:06.234
down to

234
00:06:06.679 --> 00:06:08.961
9.9%. Lower volatility and higher returns.

235
00:06:09.000 --> 00:06:09.140
Yep,

236
00:06:09.562 --> 00:06:12.398
which means the Sharpe ratio got another significant boost,

237
00:06:12.695 --> 00:06:14.148
up to 1.41.

238
00:06:14.226 --> 00:06:14.726
Very nice.

239
00:06:15.070 --> 00:06:16.007
And the maximum drawdown?

240
00:06:16.132 --> 00:06:16.726
Improved again,

241
00:06:16.929 --> 00:06:17.570
quite dramatically,

242
00:06:17.632 --> 00:06:17.929
actually,

243
00:06:17.992 --> 00:06:20.164
down to negative 14.8%.

244
00:06:20.165 --> 00:06:20.976
Under 15%

245
00:06:21.054 --> 00:06:21.476
drawdown.

246
00:06:22.257 --> 00:06:25.757
That's starting to look really compelling from a risk management perspective for investors.

247
00:06:25.878 --> 00:06:26.357
Absolutely.

248
00:06:26.458 --> 00:06:31.943
It really highlights how layering these different adaptive factors can potentially build on each other.

249
00:06:32.239 --> 00:06:32.458
Okay.

250
00:06:32.575 --> 00:06:36.802
That brings us to the final and perhaps most complex strategy they tested.

251
00:06:37.341 --> 00:06:39.364
Top five by six month momentum,

252
00:06:39.739 --> 00:06:40.583
minimum variance.

253
00:06:40.755 --> 00:06:40.974
Right.

254
00:06:41.177 --> 00:06:43.286
This one adds another layer of sophistication.

255
00:06:43.599 --> 00:06:45.943
Still starts with the top five momentum selection.

256
00:06:46.146 --> 00:06:46.443
Yes.

257
00:06:46.739 --> 00:06:49.380
Find the top five performers over the last six months.

258
00:06:50.036 --> 00:06:51.458
But then instead of just using

259
00:06:51.705 --> 00:06:53.265
inverse volatility weighting.

260
00:06:53.385 --> 00:06:54.146
It does something else.

261
00:06:54.306 --> 00:06:56.566
It uses a minimum variance optimization.

262
00:06:56.828 --> 00:07:05.785
It looks at those five chosen assets and calculates the specific weights for each one that would have resulted in the lowest possible overall portfolio volatility.

263
00:07:06.089 --> 00:07:06.230
Ah,

264
00:07:06.652 --> 00:07:09.136
so it considers not just individual asset volatility,

265
00:07:09.527 --> 00:07:11.871
but also how they move together their correlations.

266
00:07:12.042 --> 00:07:12.527
Exactly.

267
00:07:12.605 --> 00:07:13.542
That's the key difference.

268
00:07:13.933 --> 00:07:20.605
Minimum variance takes those correlations into account to try and build the smoothest possible portfolio from those top momentum assets.

269
00:07:20.753 --> 00:07:24.236
Trying to maximize diversification benefits within the selected group.

270
00:07:24.617 --> 00:07:25.076
You got it.

271
00:07:25.357 --> 00:07:26.638
It's a more holistic view of...

272
00:07:26.740 --> 00:07:27.599
They did a few things.

273
00:07:27.920 --> 00:07:28.302
First,

274
00:07:28.459 --> 00:07:30.685
they looked at returns above the risk-free rate,

275
00:07:30.920 --> 00:07:35.349
subtracting the 10-year treasury yield to see the actual risk premium earned.

276
00:07:35.584 --> 00:07:36.045
Makes sense.

277
00:07:36.046 --> 00:07:37.662
You want to know what you're getting for taking the risk.

278
00:07:37.787 --> 00:07:38.068
Right.

279
00:07:38.724 --> 00:07:41.380
Then they factored in estimated trading frictions,

280
00:07:41.427 --> 00:07:44.490
things like commissions and potential slippage when you buy or sell.

281
00:07:45.037 --> 00:07:47.099
They assumed higher costs earlier in the period,

282
00:07:47.177 --> 00:07:49.318
scaling down to about 0.5%.

283
00:07:49.380 --> 00:07:49.818
percent

284
00:07:50.005 --> 00:07:50.566
more recently.

285
00:07:50.746 --> 00:07:51.267
Realistic.

286
00:07:51.487 --> 00:07:52.366
Trading isn't free.

287
00:07:52.648 --> 00:07:53.249
And finally,

288
00:07:53.288 --> 00:07:55.710
they also deducted a hypothetical 1%

289
00:07:56.030 --> 00:07:57.093
annual management fee.

290
00:07:57.351 --> 00:07:57.573
Okay,

291
00:07:57.652 --> 00:08:01.519
so looking at that top strategy momentum plus minimum variance,

292
00:08:01.917 --> 00:08:04.456
how did it hold up after subtracting the risk-free rate,

293
00:08:04.722 --> 00:08:05.644
trading costs,

294
00:08:05.683 --> 00:08:06.175
and fees?

295
00:08:06.425 --> 00:08:09.269
So the nominal return was 15.1%.

296
00:08:09.409 --> 00:08:13.612
Subtracting the 10-year treasury yield brought that down to 10.2%

297
00:08:13.644 --> 00:08:14.816
annualized excess return.

298
00:08:15.003 --> 00:08:15.769
Still pretty solid.

299
00:08:15.956 --> 00:08:16.378
Definitely.

300
00:08:16.745 --> 00:08:16.965
Then,

301
00:08:17.045 --> 00:08:19.748
subtracting the yield and those estimated frictions and fees,

302
00:08:19.947 --> 00:08:23.350
the annualized excess return came down to 7.7%.

303
00:08:23.650 --> 00:08:24.174
7.7%

304
00:08:24.252 --> 00:08:27.756
above the risk-free rate after costs and volatility.

305
00:08:27.975 --> 00:08:32.764
The volatility stayed the same in their calculation across these adjustments at 9.4%.

306
00:08:32.826 --> 00:08:34.920
So how did the risk-adjusted performance,

307
00:08:34.921 --> 00:08:35.686
the Sharpe ratio,

308
00:08:35.826 --> 00:08:36.920
look after these deductions?

309
00:08:37.186 --> 00:08:39.748
The Sharpe was 1.61 on the nominal return.

310
00:08:40.170 --> 00:08:42.998
It dropped to 1.09 after subtracting the yield.

311
00:08:43.045 --> 00:08:43.373
And then.

312
00:08:43.589 --> 00:08:46.952
to 0.82 after subtracting the yield and the costs and fees.

313
00:08:47.013 --> 00:08:47.132
OK,

314
00:08:47.133 --> 00:08:48.015
0.82,

315
00:08:48.194 --> 00:08:49.753
still a respectable sharp ratio,

316
00:08:50.116 --> 00:08:53.941
especially considering it's after costs and represents return over treasuries.

317
00:08:54.019 --> 00:08:54.480
Exactly.

318
00:08:54.737 --> 00:08:57.159
And the maximum drawdown figures also shifted slightly.

319
00:08:57.620 --> 00:08:59.323
It was anise at 8.8 percent nominal,

320
00:08:59.566 --> 00:09:02.386
then negus at 10.2 percent after subtracting the yield,

321
00:09:02.745 --> 00:09:05.308
and a negus at 11.3 percent after all deductions.

322
00:09:05.776 --> 00:09:08.370
So even after applying these real world adjustments,

323
00:09:08.386 --> 00:09:11.339
the drawdown remains quite contained around ending at 11 percent.

324
00:09:11.757 --> 00:09:14.160
And the risk-adjusted excess returns are still compelling.

325
00:09:14.279 --> 00:09:15.381
That's the conclusion,

326
00:09:15.441 --> 00:09:15.679
really.

327
00:09:15.902 --> 00:09:17.664
Even when you account for the practicalities,

328
00:09:17.683 --> 00:09:18.683
the adaptive approach,

329
00:09:18.722 --> 00:09:20.047
especially that final strategy,

330
00:09:20.187 --> 00:09:22.711
still looks very strong compared to the passive baseline.

331
00:09:22.867 --> 00:09:23.390
And remember,

332
00:09:23.633 --> 00:09:25.008
that 7.7%

333
00:09:25.070 --> 00:09:27.109
is the return over what you'd get from treasuries.

334
00:09:27.531 --> 00:09:29.289
The paper noted the 10-year yield was about

335
00:09:29.633 --> 00:09:30.461
2.2% then,

336
00:09:30.492 --> 00:09:32.945
so your total nominal return could still be substantial.

337
00:09:33.133 --> 00:09:33.320
Right.

338
00:09:33.336 --> 00:09:34.539
So wrapping this up,

339
00:09:35.039 --> 00:09:38.367
what's the main message for someone listening thinking about their own allocation?

340
00:09:38.529 --> 00:09:44.735
I think the key takeaway is that this paper really demonstrates quite clearly through backtesting how adaptive strategies,

341
00:09:44.856 --> 00:09:48.641
ones that react to market conditions using things like momentum and volatility,

342
00:09:49.219 --> 00:09:52.282
have the potential to significantly improve risk adjusted returns.

343
00:09:52.664 --> 00:09:54.422
Compared to just holding a static mix.

344
00:09:54.782 --> 00:09:55.047
Yes,

345
00:09:55.532 --> 00:10:03.094
especially strategies that combine factors like using momentum to select assets and then using volatility or minimum variance techniques to construct the portfolio.

346
00:10:03.735 --> 00:10:08.032
They seem to offer better returns and substantially reduced drawdowns in the periods studied.

347
00:10:08.393 --> 00:10:10.986
A compelling argument for not just setting and forgetting,

348
00:10:11.590 --> 00:10:14.213
but adapting based on what the market is actually doing.

349
00:10:14.293 --> 00:10:14.893
Precisely.

350
00:10:14.992 --> 00:10:18.055
It's about dynamically managing both risk and return potential.

351
00:10:18.277 --> 00:10:20.699
Thank you for tuning in to Papers with Backtest podcast.

352
00:10:21.199 --> 00:10:23.340
We hope today's episode gave you useful insights.

353
00:10:23.902 --> 00:10:25.863
Join us next time as we break down more research.

354
00:10:26.809 --> 00:10:28.199
And for more papers and backtests,

355
00:10:28.434 --> 00:10:31.574
find us at https.paperswithbacktest.com.

356
00:10:32.309 --> 00:10:32.934
Happy trading.