WEBVTT

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

2
00:00:00.280 --> 00:00:02.162
welcome back to Papers of Backtest podcast.

3
00:00:02.402 --> 00:00:04.785
Today we dive into another algo trading research paper.

4
00:00:04.906 --> 00:00:05.305
We are,

5
00:00:05.324 --> 00:00:08.445
and today we're zeroing in on a study called Analyst Days,

6
00:00:08.789 --> 00:00:09.633
Stock Prices,

7
00:00:09.711 --> 00:00:10.891
and Firm Performance.

8
00:00:11.531 --> 00:00:13.914
It's by Diwu and Amir Yarin.

9
00:00:14.055 --> 00:00:14.250
Right,

10
00:00:14.453 --> 00:00:16.359
from October 2018.

11
00:00:16.360 --> 00:00:16.438
Yeah.

12
00:00:16.439 --> 00:00:18.797
And it digs into these things called analyst days.

13
00:00:19.031 --> 00:00:20.016
What exactly are those,

14
00:00:20.375 --> 00:00:20.953
based on the paper?

15
00:00:21.047 --> 00:00:21.188
Well,

16
00:00:21.359 --> 00:00:23.438
they're basically events hosted by a company.

17
00:00:23.609 --> 00:00:25.484
They invite equity analysts,

18
00:00:25.562 --> 00:00:26.906
institutional investors,

19
00:00:27.641 --> 00:00:27.891
you know,

20
00:00:28.094 --> 00:00:28.844
the big players.

21
00:00:29.016 --> 00:00:29.125
Okay.

22
00:00:29.380 --> 00:00:30.441
And they share information,

23
00:00:30.741 --> 00:00:31.182
updates,

24
00:00:31.262 --> 00:00:31.663
plans,

25
00:00:31.682 --> 00:00:32.303
that sort of thing.

26
00:00:32.362 --> 00:00:35.788
But the key is because of regulation fair disclosure or a reg FD.

27
00:00:35.827 --> 00:00:35.968
Oh,

28
00:00:36.108 --> 00:00:36.448
reg FD.

29
00:00:36.667 --> 00:00:37.167
Exactly.

30
00:00:37.546 --> 00:00:41.210
Anything significant they disclose has to be released publicly at the same time.

31
00:00:41.249 --> 00:00:42.913
So no selective disclosure.

32
00:00:43.296 --> 00:00:43.632
Got it.

33
00:00:43.633 --> 00:00:45.233
So it's a burst of public information.

34
00:00:45.639 --> 00:00:48.639
And the paper says these have become more popular since around 2004.

35
00:00:48.905 --> 00:00:49.061
Yeah,

36
00:00:49.093 --> 00:00:50.343
they've seen a rise in usage.

37
00:00:50.436 --> 00:00:51.968
Firms seem to find them valuable.

38
00:00:52.218 --> 00:00:52.436
Okay.

39
00:00:52.499 --> 00:00:58.202
So our mission today then is to unpack the trading rules and the backtest results the researchers found.

40
00:00:58.544 --> 00:00:58.844
Basically,

41
00:00:58.884 --> 00:01:01.026
how does the market react after these analyst days?

42
00:01:01.487 --> 00:01:01.987
Exactly.

43
00:01:02.187 --> 00:01:03.808
What happens to the stock price?

44
00:01:03.948 --> 00:01:06.273
And could you potentially treat on that information?

45
00:01:06.433 --> 00:01:08.214
And they had a good amount of data to work with,

46
00:01:08.253 --> 00:01:08.675
I gather.

47
00:01:08.870 --> 00:01:08.995
Oh,

48
00:01:09.058 --> 00:01:09.214
yeah.

49
00:01:09.355 --> 00:01:10.355
Quite comprehensive.

50
00:01:10.737 --> 00:01:15.245
They looked at 3,890 analyst day events.

51
00:01:15.401 --> 00:01:15.620
Wow.

52
00:01:15.823 --> 00:01:16.089
U.S.

53
00:01:16.167 --> 00:01:19.792
listed firms between 2004 and 2015.

54
00:01:19.793 --> 00:01:23.261
And they linked all that event data with stock prices,

55
00:01:23.370 --> 00:01:24.230
accounting data.

56
00:01:24.636 --> 00:01:25.292
Really thorough.

57
00:01:25.504 --> 00:01:29.569
And they even collected the text from the announcement so they could see what was being discussed.

58
00:01:29.570 --> 00:01:29.768
Okay,

59
00:01:29.869 --> 00:01:30.729
great foundation.

60
00:01:31.069 --> 00:01:31.248
Yeah.

61
00:01:31.529 --> 00:01:33.393
Let's get straight to the core findings then.

62
00:01:34.213 --> 00:01:36.752
What happened to stock prices after these analyst days?

63
00:01:37.252 --> 00:01:38.776
Were there any signals for traders?

64
00:01:39.518 --> 00:01:40.338
The main takeaway,

65
00:01:40.721 --> 00:01:42.174
firms holding these events saw,

66
00:01:42.502 --> 00:01:43.205
on average,

67
00:01:43.440 --> 00:01:46.643
significantly higher abnormal returns afterwards.

68
00:01:46.768 --> 00:01:48.580
Abnormal returns meaning they beat the market,

69
00:01:48.627 --> 00:01:48.846
right?

70
00:01:48.940 --> 00:01:49.502
Precisely.

71
00:01:49.752 --> 00:01:54.737
They performed better than you'd expect just based on general market trends or their typical risk profile.

72
00:01:54.824 --> 00:01:55.665
Can we put a number on that?

73
00:01:55.705 --> 00:01:56.245
How much better?

74
00:01:56.346 --> 00:01:56.485
Well,

75
00:01:56.506 --> 00:01:58.807
one of the first back tests they ran was pretty simple.

76
00:01:59.409 --> 00:02:00.710
Buy the stock on the analyst day,

77
00:02:01.112 --> 00:02:02.530
hold it for 20 trading days.

78
00:02:02.632 --> 00:02:02.831
Okay.

79
00:02:02.893 --> 00:02:03.690
Straightforward enough.

80
00:02:03.995 --> 00:02:07.073
That strategy earned a market adjusted return of 1.6%.

81
00:02:07.979 --> 00:02:09.456
So 1.6%

82
00:02:09.457 --> 00:02:12.065
above the market over those 20 days on average.

83
00:02:12.706 --> 00:02:13.175
1.6%

84
00:02:13.456 --> 00:02:14.050
in about a month.

85
00:02:14.065 --> 00:02:14.300
That's,

86
00:02:15.128 --> 00:02:15.237
whoa,

87
00:02:15.518 --> 00:02:16.409
that's certainly noticeable.

88
00:02:16.410 --> 00:02:18.300
Did they look at it from a portfolio perspective too?

89
00:02:18.425 --> 00:02:18.909
They did.

90
00:02:19.308 --> 00:02:22.949
They simulated building a portfolio of these stocks around their analyst days,

91
00:02:22.968 --> 00:02:24.271
a calendar time approach.

92
00:02:24.951 --> 00:02:29.830
That portfolio showed a one-month four-factor alpha of 1.8%.

93
00:02:29.908 --> 00:02:30.126
Okay,

94
00:02:30.291 --> 00:02:30.533
alpha.

95
00:02:30.634 --> 00:02:33.455
So that's return adjusted for market risk,

96
00:02:33.572 --> 00:02:34.009
size,

97
00:02:34.072 --> 00:02:34.455
value,

98
00:02:34.494 --> 00:02:34.955
momentum.

99
00:02:35.212 --> 00:02:35.384
Yeah.

100
00:02:35.666 --> 00:02:36.712
The usual factors.

101
00:02:36.806 --> 00:02:37.275
Exactly.

102
00:02:37.619 --> 00:02:40.259
It suggests there was some genuine excess return there,

103
00:02:40.306 --> 00:02:42.431
not just compensation for standard risk factors.

104
00:02:42.432 --> 00:02:42.619
Right.

105
00:02:42.837 --> 00:02:43.978
And 1.8%

106
00:02:44.041 --> 00:02:45.978
alpha per month is quite substantial.

107
00:02:46.119 --> 00:02:47.353
And was this just a quick blip?

108
00:02:47.720 --> 00:02:48.421
Or did it last?

109
00:02:48.461 --> 00:02:49.361
That's the interesting part.

110
00:02:49.382 --> 00:02:50.744
It wasn't just a flash in the pan.

111
00:02:51.043 --> 00:02:54.568
The abnormal returns stayed significantly positive for up to six months.

112
00:02:54.685 --> 00:02:55.166
Six months.

113
00:02:55.228 --> 00:02:55.486
Wow.

114
00:02:55.588 --> 00:02:55.768
Yeah.

115
00:02:55.971 --> 00:03:01.830
And they didn't find evidence that this was because the stock suddenly got riskier or that it was just some kind of temporary bounce back,

116
00:03:01.853 --> 00:03:02.072
you know,

117
00:03:02.135 --> 00:03:02.893
mean reversion.

118
00:03:02.994 --> 00:03:05.924
So no obvious increase in risk and the gains held for a while.

119
00:03:06.236 --> 00:03:06.502
Right.

120
00:03:07.018 --> 00:03:07.158
Now,

121
00:03:07.221 --> 00:03:07.705
eventually,

122
00:03:07.893 --> 00:03:10.064
after about 250 trading days,

123
00:03:10.268 --> 00:03:11.080
roughly a year,

124
00:03:11.783 --> 00:03:14.486
the returns weren't statistically significant anymore.

125
00:03:15.104 --> 00:03:17.285
But they didn't see a big downward trend either.

126
00:03:17.326 --> 00:03:18.488
It just sort of faded.

127
00:03:18.809 --> 00:03:19.910
That implies the market is,

128
00:03:20.949 --> 00:03:21.129
well,

129
00:03:21.449 --> 00:03:22.492
maybe a bit slow to catch on.

130
00:03:22.731 --> 00:03:25.152
It takes months to fully price in the information.

131
00:03:25.473 --> 00:03:27.379
That's exactly what the researchers suggest.

132
00:03:27.418 --> 00:03:29.160
They talk about market underreaction.

133
00:03:29.582 --> 00:03:31.301
The cumulative abnormal returns,

134
00:03:31.348 --> 00:03:32.098
the SARs,

135
00:03:32.457 --> 00:03:34.941
kept trending upwards for the first three months.

136
00:03:35.098 --> 00:03:36.145
And trading volume,

137
00:03:36.285 --> 00:03:36.910
did that change?

138
00:03:36.941 --> 00:03:37.301
It did.

139
00:03:37.441 --> 00:03:40.582
Volume was higher in the first 20 trading days post-event,

140
00:03:41.145 --> 00:03:41.426
which,

141
00:03:41.754 --> 00:03:42.113
again,

142
00:03:42.223 --> 00:03:43.176
fits that picture.

143
00:03:43.448 --> 00:03:46.572
of investors gradually digesting and reacting to the news.

144
00:03:46.871 --> 00:03:47.031
Okay,

145
00:03:47.032 --> 00:03:49.135
so the action seems to be after the event.

146
00:03:49.834 --> 00:03:51.275
What about trying to get in before?

147
00:03:52.174 --> 00:03:53.838
Did the market anticipate these events?

148
00:03:54.158 --> 00:03:54.978
Apparently not,

149
00:03:55.002 --> 00:03:56.080
or at least not consistently.

150
00:03:56.299 --> 00:04:01.252
The abnormal returns in the month leading up to the analyst day weren't significantly different from zero.

151
00:04:01.642 --> 00:04:06.049
So just knowing an analyst day was scheduled didn't really help you from a trading perspective beforehand?

152
00:04:06.236 --> 00:04:07.799
Doesn't look like it based on their data.

153
00:04:08.049 --> 00:04:11.470
The value seems to be in the information revealed on the day itself.

154
00:04:11.799 --> 00:04:12.299
which means

155
00:04:12.472 --> 00:04:14.914
the opportunity window opens right around the event,

156
00:04:15.075 --> 00:04:16.215
or maybe even slightly after.

157
00:04:16.596 --> 00:04:17.698
That seems to be the case.

158
00:04:18.057 --> 00:04:21.120
They found you could realize the average 1.6%

159
00:04:21.121 --> 00:04:24.143
monthly abnormal return by buying on the analyst day.

160
00:04:24.807 --> 00:04:27.463
But even buying up to a week later still seemed profitable.

161
00:04:27.885 --> 00:04:28.635
The SARA-RR,

162
00:04:28.713 --> 00:04:33.666
the cumulative return from day T plus 5 to day T plus 20 was still around 1%.

163
00:04:34.213 --> 00:04:37.541
So you didn't necessarily need lightning fast reactions.

164
00:04:37.573 --> 00:04:38.370
There was some time.

165
00:04:38.557 --> 00:04:39.276
A little bit of time,

166
00:04:39.338 --> 00:04:39.557
yes.

167
00:04:40.044 --> 00:04:42.638
The under reaction created a slightly longer window.

168
00:04:43.065 --> 00:04:45.388
And they check this using different ways to measure returns,

169
00:04:45.408 --> 00:04:45.968
just to be sure.

170
00:04:46.349 --> 00:04:47.248
Robustness checks.

171
00:04:47.451 --> 00:04:47.849
Absolutely.

172
00:04:48.029 --> 00:04:48.548
They used

173
00:04:49.373 --> 00:04:51.377
Carhartt four-factor model returns,

174
00:04:51.712 --> 00:04:53.337
buy and hold abnormal returns,

175
00:04:53.431 --> 00:04:54.158
BHRs,

176
00:04:54.580 --> 00:04:56.337
looked at various windows like T20 to

177
00:04:56.822 --> 00:04:57.658
T plus 60.

178
00:04:57.861 --> 00:04:58.720
And the results held up.

179
00:04:59.041 --> 00:04:59.462
Largely,

180
00:04:59.587 --> 00:04:59.869
yes.

181
00:05:00.087 --> 00:05:01.837
Similar positive trends across the board.

182
00:05:02.056 --> 00:05:07.306
They did note maybe a tiny hint of mean reversion in one BHR measure between 30 and 60 days,

183
00:05:07.353 --> 00:05:09.134
but the standard errors were pretty high,

184
00:05:09.197 --> 00:05:10.728
so not a strong signal.

185
00:05:10.994 --> 00:05:11.150
Okay,

186
00:05:11.151 --> 00:05:12.775
so the short to medium term picture looks...

187
00:05:12.817 --> 00:05:13.418
It's pretty solid.

188
00:05:13.878 --> 00:05:16.903
What about even longer term using that calendar portfolio approach?

189
00:05:17.243 --> 00:05:17.422
Yeah,

190
00:05:17.501 --> 00:05:18.282
looking out further,

191
00:05:18.383 --> 00:05:19.563
up to 60 months,

192
00:05:19.665 --> 00:05:25.606
it confirmed those strong positive abnormal returns on the day and for up to six months after.

193
00:05:25.809 --> 00:05:26.028
Right.

194
00:05:26.309 --> 00:05:30.051
They calculated annualized alphas reaching as high as 8%

195
00:05:30.380 --> 00:05:32.161
in those first two months post-event.

196
00:05:32.380 --> 00:05:32.958
8%

197
00:05:33.083 --> 00:05:33.833
annualized alpha.

198
00:05:34.083 --> 00:05:34.755
That's significant.

199
00:05:34.895 --> 00:05:35.395
It is.

200
00:05:35.442 --> 00:05:35.801
And again,

201
00:05:35.817 --> 00:05:36.505
they checked the risk.

202
00:05:36.567 --> 00:05:40.067
The post-event betas weren't significantly different from pre-event betas.

203
00:05:40.239 --> 00:05:42.567
So it wasn't just taking on more risk.

204
00:05:42.753 --> 00:05:43.834
And beyond six months.

205
00:05:44.175 --> 00:05:46.556
The alpha tended to drift back towards zero,

206
00:05:46.778 --> 00:05:49.919
especially when you held for more than 250 trading days.

207
00:05:50.681 --> 00:05:50.798
Oh,

208
00:05:50.821 --> 00:05:52.103
and an interesting side note.

209
00:05:53.040 --> 00:05:56.611
These stocks tended to have negative commovement with momentum stocks.

210
00:05:57.025 --> 00:05:59.673
Meaning when momentum strategies were hot,

211
00:06:00.251 --> 00:06:03.329
these analyst day stocks maybe didn't do as well and vice versa.

212
00:06:03.407 --> 00:06:03.814
Kind of,

213
00:06:03.861 --> 00:06:04.017
yeah.

214
00:06:04.267 --> 00:06:07.001
An interesting portfolio diversification aspect,

215
00:06:07.064 --> 00:06:07.423
perhaps.

216
00:06:07.454 --> 00:06:09.267
Let's circle back to the trading volume patterns.

217
00:06:09.423 --> 00:06:09.579
Yeah.

218
00:06:09.595 --> 00:06:11.079
You said it picked up after the event.

219
00:06:11.220 --> 00:06:11.439
Right.

220
00:06:11.697 --> 00:06:12.857
Not much happening before,

221
00:06:12.957 --> 00:06:15.058
maybe a slight pickup a day or two prior.

222
00:06:15.576 --> 00:06:15.978
But then,

223
00:06:16.197 --> 00:06:17.197
bang on the event day,

224
00:06:17.357 --> 00:06:18.857
volume jumps about 7%

225
00:06:18.939 --> 00:06:19.517
above normal.

226
00:06:19.857 --> 00:06:20.482
The day after.

227
00:06:20.920 --> 00:06:21.420
Even higher,

228
00:06:21.537 --> 00:06:22.060
about 9%

229
00:06:22.303 --> 00:06:22.842
above normal.

230
00:06:23.123 --> 00:06:26.318
And this elevated volume stuck around for about 20 trading days,

231
00:06:26.599 --> 00:06:27.264
averaging maybe

232
00:06:27.928 --> 00:06:30.506
3.4% higher overall during that period.

233
00:06:30.678 --> 00:06:35.318
That really paints a picture of information slowly diffusing and triggering trades,

234
00:06:35.349 --> 00:06:35.756
doesn't it?

235
00:06:36.146 --> 00:06:37.490
Supports the underreaction idea.

236
00:06:37.678 --> 00:06:38.396
It really does.

237
00:06:38.537 --> 00:06:39.303
Increased interest.

238
00:06:39.545 --> 00:06:41.327
More trading as the news sinks in.

239
00:06:41.728 --> 00:06:41.888
Now,

240
00:06:42.269 --> 00:06:45.249
did the type of information shared at these analyst days matter?

241
00:06:45.452 --> 00:06:46.952
Did they all have the same effect?

242
00:06:47.171 --> 00:06:47.394
Ah,

243
00:06:47.612 --> 00:06:48.152
good question.

244
00:06:48.456 --> 00:06:51.097
They actually looked into that using the text analysis.

245
00:06:51.214 --> 00:06:52.175
They use a technique,

246
00:06:52.878 --> 00:06:54.456
latent Dirichlet allocation,

247
00:06:54.941 --> 00:06:55.284
LDA.

248
00:06:55.503 --> 00:06:55.675
Okay,

249
00:06:55.816 --> 00:06:57.347
fancy term for grouping topics.

250
00:06:57.566 --> 00:06:57.987
Basically,

251
00:06:58.050 --> 00:06:58.222
yeah.

252
00:06:58.909 --> 00:07:01.206
They classified the announcements into four main types.

253
00:07:01.550 --> 00:07:02.441
Product announcements,

254
00:07:02.581 --> 00:07:03.847
reviews of past results,

255
00:07:04.253 --> 00:07:05.612
discussions about strategy,

256
00:07:05.816 --> 00:07:07.487
and talks about technology and markets.

257
00:07:07.816 --> 00:07:09.097
And did the market react differently?

258
00:07:09.297 --> 00:07:09.878
Very much so.

259
00:07:10.538 --> 00:07:15.002
The big winners were analyst days focused on new products or discussing technology and markets.

260
00:07:15.041 --> 00:07:15.885
How much better were they?

261
00:07:16.123 --> 00:07:16.264
Well,

262
00:07:16.322 --> 00:07:19.186
product-related days showed suckars up to 8%

263
00:07:19.209 --> 00:07:20.631
in the two months following the event.

264
00:07:20.686 --> 00:07:21.287
That's huge.

265
00:07:21.451 --> 00:07:22.006
8%?

266
00:07:22.147 --> 00:07:22.287
Yeah.

267
00:07:22.514 --> 00:07:22.912
Compared to?

268
00:07:23.053 --> 00:07:25.967
Compared to events just reviewing past financial results.

269
00:07:26.451 --> 00:07:29.358
Those didn't generate significantly positive returns at all.

270
00:07:29.717 --> 00:07:30.170
Interesting.

271
00:07:30.201 --> 00:07:31.873
So forward-looking news,

272
00:07:32.420 --> 00:07:33.639
like products and markets,

273
00:07:33.983 --> 00:07:37.279
had a much bigger impact than just rehashing old numbers.

274
00:07:37.420 --> 00:07:37.967
Exactly.

275
00:07:38.501 --> 00:07:41.184
Strategy-related events also showed a positive drift,

276
00:07:41.303 --> 00:07:42.826
maybe 1.6%

277
00:07:42.827 --> 00:07:44.248
CAR in the first 20 days.

278
00:07:44.768 --> 00:07:48.170
And the technology-market-focused ones had cars up to 4%

279
00:07:48.568 --> 00:07:49.451
over 60 days.

280
00:07:50.295 --> 00:07:52.654
But product news was the clear standout.

281
00:07:53.076 --> 00:07:54.139
That makes intuitive sense,

282
00:07:54.178 --> 00:07:54.693
I suppose.

283
00:07:54.756 --> 00:07:56.865
New products often mean new growth potential.

284
00:07:57.053 --> 00:07:57.522
Precisely.

285
00:07:57.678 --> 00:07:58.943
It's about future prospects.

286
00:07:59.162 --> 00:08:03.912
The paper also touched briefly on actual firm performance and analyst behavior after these events,

287
00:08:03.990 --> 00:08:04.193
right?

288
00:08:04.225 --> 00:08:05.084
Not just stock price?

289
00:08:05.131 --> 00:08:05.490
It did,

290
00:08:05.522 --> 00:08:05.662
yeah.

291
00:08:06.173 --> 00:08:06.734
Just briefly,

292
00:08:06.814 --> 00:08:10.558
they found firms holding these days tended to show significantly higher revenue growth,

293
00:08:10.898 --> 00:08:11.917
better earnings per share,

294
00:08:11.999 --> 00:08:14.679
and higher dividend yields for up to two years afterwards.

295
00:08:14.859 --> 00:08:18.421
So the events often preceded actual improvements in the business itself?

296
00:08:18.523 --> 00:08:19.570
It seems correlated,

297
00:08:19.624 --> 00:08:19.890
yes.

298
00:08:20.163 --> 00:08:21.648
And analyst behavior changed too.

299
00:08:22.155 --> 00:08:23.171
After an analyst day,

300
00:08:23.249 --> 00:08:25.624
you typically saw more analysts covering the stock.

301
00:08:25.765 --> 00:08:26.452
Increased coverage.

302
00:08:26.796 --> 00:08:27.921
Higher earnings estimates,

303
00:08:28.312 --> 00:08:29.390
higher price targets,

304
00:08:29.734 --> 00:08:30.499
and importantly,

305
00:08:30.937 --> 00:08:32.780
less disagreement among analysts,

306
00:08:33.015 --> 00:08:34.296
lower forecast dispersion.

307
00:08:34.777 --> 00:08:38.942
So the events seem to provide clarity and maybe some optimism to the analyst community.

308
00:08:39.121 --> 00:08:40.223
That seems to be the effect.

309
00:08:40.301 --> 00:08:40.524
Yes,

310
00:08:40.621 --> 00:08:41.383
more coverage,

311
00:08:41.504 --> 00:08:42.184
more agreement,

312
00:08:42.344 --> 00:08:43.606
more positive outlooks.

313
00:08:43.762 --> 00:08:44.004
Okay,

314
00:08:44.043 --> 00:08:47.567
so let's wrap this up for someone listening who might be interested in the trading angle.

315
00:08:48.207 --> 00:08:49.129
What's the main takeaway?

316
00:08:49.348 --> 00:08:55.653
The core takeaway is that this study reveals significantly positive abnormal returns following analyst days.

317
00:08:56.184 --> 00:08:58.012
There seems to be a market underreaction.

318
00:08:58.074 --> 00:08:59.621
Meaning there's potentially an edge there.

319
00:09:00.168 --> 00:09:00.621
Potentially,

320
00:09:00.746 --> 00:09:00.965
yes.

321
00:09:01.269 --> 00:09:06.450
especially for analyst days focused on those forward-looking topics like product announcements or market updates.

322
00:09:06.771 --> 00:09:08.669
The opportunity might persist for days,

323
00:09:08.829 --> 00:09:09.169
weeks,

324
00:09:09.435 --> 00:09:10.872
even a few months after the event.

325
00:09:10.989 --> 00:09:13.114
But and this is the crucial caveat.

326
00:09:13.372 --> 00:09:13.810
Absolutely.

327
00:09:13.849 --> 00:09:17.591
This is based on historical averages from 2004 to 2015.

328
00:09:17.638 --> 00:09:18.435
Markets change.

329
00:09:18.794 --> 00:09:21.716
Any specific strategy needs serious individual research,

330
00:09:22.028 --> 00:09:23.481
backtesting on current data.

331
00:09:23.872 --> 00:09:24.450
You know the drill.

332
00:09:24.716 --> 00:09:26.060
Past performance is no guarantee,

333
00:09:26.122 --> 00:09:26.481
et cetera,

334
00:09:26.482 --> 00:09:26.888
et cetera.

335
00:09:27.294 --> 00:09:30.138
But it definitely highlights analyst days as events.

336
00:09:30.338 --> 00:09:31.377
worth paying attention to.

337
00:09:31.843 --> 00:09:32.223
Definitely.

338
00:09:32.323 --> 00:09:34.383
It's a specific type of information event that,

339
00:09:34.783 --> 00:09:35.744
historically at least,

340
00:09:35.842 --> 00:09:37.744
seems to have had a predictable,

341
00:09:37.885 --> 00:09:38.885
albeit delayed,

342
00:09:38.983 --> 00:09:39.725
market impact.

343
00:09:39.881 --> 00:09:40.686
Fascinating stuff.

344
00:09:40.905 --> 00:09:43.389
A clear example of potential market inefficiency,

345
00:09:43.865 --> 00:09:44.764
even if it's temporary.

346
00:09:44.944 --> 00:09:45.389
Indeed.

347
00:09:45.764 --> 00:09:50.631
Makes you wonder what other scheduled information events might have similar underappreciated effects.

348
00:09:51.381 --> 00:09:53.569
Thank you for tuning in to Papers with Backtest podcast.

349
00:09:53.944 --> 00:09:56.069
We hope today's episode gave you useful insights.

350
00:09:56.522 --> 00:09:58.615
Join us next time as we break down more research.

351
00:09:59.069 --> 00:10:00.365
And for more papers and backtests,

352
00:10:00.551 --> 00:10:03.616
find us at https.paperswithbagtest.com.

353
00:10:04.157 --> 00:10:04.839
Happy trading!