My understanding is that for the vast majority of companies, P/B is a poor valuation metric. The main exception would be companies whose main assets are marked to market like cash or marketable securities (e.g. shares of stock).
For companies with steady trending earnings, P/E should be good. Otherwise P/S may be better with the main caveat that you have reason to believe that their profit margin will mean revert which will not always be true.
I wish I knew. I can’t get it to replicate either. Immediately after my post I tried to reply again but there was no message option this time.
I remember Jim mentioned:
P/B: for company mostly investing it’s fund
P/E: for company which has steady/stable/long earning history
Hope I didn’t get it wrong.
@cwags02
I’ve got my data analysis pipeline going in R, and ran on the free price to book data for BRK (and other companies) that I pulled down. Later, I can go to library and get M* data or maybe try a fee-based service, but if it’s trivial I was wondering how the free PriceToBook data I pulled down for BRK compares to your M* data?
If it’s not trivial pls don’t waste time 
Date,BRKPriceToBook
2011-02-28,1.3742
2011-03-31,1.2912
2011-04-30,1.2858
2011-05-31,1.2228
2011-06-30,1.1743
2011-07-31,1.1268
2011-08-31,1.1108
2011-09-30,1.1015
2011-10-31,1.2066
2011-11-30,1.2213
2011-12-31,1.1478
2012-01-31,1.1793
2012-02-29,1.1799
2012-03-31,1.1427
2012-04-30,1.1326
2012-05-31,1.1159
2012-06-30,1.1633
2012-07-31,1.1857
2012-08-31,1.1786
2012-09-30,1.186
2012-10-31,1.1591
2012-11-30,1.1817
2012-12-31,1.1825
2013-01-31,1.2828
2013-02-28,1.3464
2013-03-31,1.2966
2013-04-30,1.3209
2013-05-31,1.4206
2013-06-30,1.3689
2013-07-31,1.4144
2013-08-31,1.3584
2013-09-30,1.3437
2013-10-31,1.3633
2013-11-30,1.3787
2013-12-31,1.3178
2014-01-31,1.2484
2014-02-28,1.2868
2014-03-31,1.3536
2014-04-30,1.396
2014-05-31,1.3888
2014-06-30,1.3342
2014-07-31,1.322
2014-08-31,1.445
2014-09-30,1.4321
2014-10-31,1.4533
2014-11-30,1.5432
2014-12-31,1.5433
2015-01-31,1.4766
2015-02-28,1.5128
2015-03-31,1.4765
2015-04-30,1.4467
2015-05-31,1.4606
2015-06-30,1.3657
2015-07-31,1.4294
2015-08-31,1.3477
2015-09-30,1.2935
2015-10-31,1.3523
2015-11-30,1.332
2015-12-31,1.2728
2016-01-31,1.2508
2016-02-29,1.2984
2016-03-31,1.3542
2016-04-30,1.3889
2016-05-31,1.3424
2016-06-30,1.3566
2016-07-31,1.3511
2016-08-31,1.4109
2016-09-30,1.3216
2016-10-31,1.3192
2016-11-30,1.4444
2016-12-31,1.4239
2017-01-31,1.4344
2017-02-28,1.4989
2017-03-31,1.4035
2017-04-30,1.3914
2017-05-31,1.3936
2017-06-30,1.3913
2017-07-31,1.4365
2017-08-31,1.4857
2017-09-30,1.4664
2017-10-31,1.4961
2017-11-30,1.5495
2017-12-31,1.4047
2018-01-31,1.5225
2018-02-28,1.4667
2018-03-31,1.4166
2018-04-30,1.3761
2018-05-31,1.3602
2018-06-30,1.2905
2018-07-31,1.3734
2018-08-31,1.4439
2018-09-30,1.4044
2018-10-31,1.3482
2018-11-30,1.4282
2018-12-31,1.4413
2019-01-31,1.4582
2019-02-28,1.4215
2019-03-31,1.3403
2019-04-30,1.4461
2019-05-31,1.3146
2019-06-30,1.3642
2019-07-31,1.3182
2019-08-31,1.2995
2019-09-30,1.282
2019-10-31,1.3106
2019-11-30,1.3548
2019-12-31,1.3034
2020-01-31,1.2907
2020-02-29,1.1829
2020-03-31,1.1947
2020-04-30,1.2299
2020-05-31,1.2142
2020-06-30,1.1016
2020-07-31,1.209
2020-08-31,1.3242
2020-09-30,1.2259
2020-10-31,1.1609
2020-11-30,1.2934
2020-12-31,1.2268
2021-01-31,1.209
2021-02-28,1.2781
2021-03-31,1.3136
2021-04-30,1.41
2021-05-31,1.478
2021-06-30,1.3515
2021-07-31,1.3529
2021-08-31,1.3749
2021-09-30,1.3085
2021-10-31,1.3763
2021-11-30,1.3103
2021-12-31,1.3219
2022-01-31,1.3814
2022-02-28,1.3988
2022-03-31,1.5373
2022-04-30,1.4068
2022-05-31,1.3721
2022-06-30,1.3052
2022-07-31,1.439
2022-08-31,1.3401
2022-09-30,1.282
2022-10-11,1.2772
@johnIII & @tedthedog - I’m really appreciating your comments & digging into @mungofitch’s work. Taking BVpeak and smoothing the price data similar to what was mentioned in the quote above yields something very interesting: inflation +33% with really pretty data.
I’m not sure how to / if I want to make the adjustments to book value like Jim did and @johnIII mentions; I don’t think I’m that smart or follow Berkshire’s holdings close enough to make an informed decision.
Image uploads still aren’t working at the fool so the tables / charts are available here:
I hope with every fiber of my being the Carmax results represent the future because holy buckets that’s good.
The carmax curve fit is a bit goofy. I used a cubic simply because we’re in the steep part of the hocky stick and the other fit options really missed. Based on how the data lay out, it seems doing two different linear fits, one in the steep hocky stick part and one in the other, would be good enough and give a similar enough answer.
We’re similar.
Graphical differences here: https://i.imgur.com/W6141QP.jpg Data are below along with a table for the most egregious offenders.
Who is right?
| Date | Tedthedog | cwags02 | Delta | |
|---|---|---|---|---|
| 2/28/2011 | 1.3742 | 1.37189 | -0.00231 | |
| 3/31/2011 | 1.2912 | 1.29217 | 0.00097 | |
| 4/30/2011 | 1.2858 | 1.28718 | 0.00138 | |
| 5/31/2011 | 1.2228 | 1.22182 | -0.00098 | |
| 6/30/2011 | 1.1743 | 1.17595 | 0.00165 | |
| 7/31/2011 | 1.1268 | 1.12703 | 0.00023 | |
| 8/31/2011 | 1.1108 | 1.10925 | -0.00155 | |
| 9/30/2011 | 1.1015 | 1.09996 | -0.00154 | |
| 10/31/2011 | 1.2066 | 1.20528 | -0.00132 | |
| 11/30/2011 | 1.2213 | 1.21921 | -0.00209 | |
| 12/31/2011 | 1.1478 | 1.18116 | 0.03336 | |
| 1/31/2012 | 1.1793 | 1.17719 | -0.00211 | |
| 2/29/2012 | 1.1799 | 1.17846 | -0.00144 | |
| 3/31/2012 | 1.1427 | 1.21924 | 0.07654 | |
| 4/30/2012 | 1.1326 | 1.1322 | -0.0004 | |
| 5/31/2012 | 1.1159 | 1.11686 | 0.00096 | |
| 6/30/2012 | 1.1633 | 1.17321 | 0.00991 | |
| 7/31/2012 | 1.1857 | 1.18529 | -0.00041 | |
| 8/31/2012 | 1.1786 | 1.17831 | -0.00029 | |
| 9/30/2012 | 1.186 | 1.23245 | 0.04645 | |
| 10/31/2012 | 1.1591 | 1.15946 | 0.00036 | |
| 11/30/2012 | 1.1817 | 1.18269 | 0.00099 | |
| 12/31/2012 | 1.1825 | 1.17805 | -0.00445 | |
| 1/31/2013 | 1.2828 | 1.27301 | -0.00979 | |
| 2/28/2013 | 1.3464 | 1.34189 | -0.00451 | |
| 3/31/2013 | 1.2966 | 1.36886 | 0.07226 | |
| 4/30/2013 | 1.3209 | 1.39678 | 0.07588 | |
| 5/31/2013 | 1.4206 | 1.41973 | -0.00087 | |
| 6/30/2013 | 1.3689 | 1.39319 | 0.02429 | |
| 7/31/2013 | 1.4144 | 1.44239 | 0.02799 | |
| 8/31/2013 | 1.3584 | 1.35746 | -0.00094 | |
| 9/30/2013 | 1.3437 | 1.38547 | 0.04177 | |
| 10/31/2013 | 1.3633 | 1.40478 | 0.04148 | |
| 11/30/2013 | 1.3787 | 1.3789 | 0.0002 | |
| 12/31/2013 | 1.3178 | 1.403 | 0.0852 | |
| 1/31/2014 | 1.2484 | 1.32064 | 0.07224 | |
| 2/28/2014 | 1.2868 | 1.37015 | 0.08335 | |
| 3/31/2014 | 1.3536 | 1.38911 | 0.03551 | |
| 4/30/2014 | 1.396 | 1.4323 | 0.0363 | |
| 5/31/2014 | 1.3888 | 1.39077 | 0.00197 | |
| 6/30/2014 | 1.3342 | 1.36979 | 0.03559 | |
| 7/31/2014 | 1.322 | 1.35758 | 0.03558 | |
| 8/31/2014 | 1.445 | 1.44493 | -7E-05 | |
| 9/30/2014 | 1.4321 | 1.4547 | 0.0226 | |
| 10/31/2014 | 1.4533 | 1.47599 | 0.02269 | |
| 11/30/2014 | 1.5432 | 1.54306 | -0.00014 | |
| 12/31/2014 | 1.5433 | 1.55828 | 0.01498 | |
| 1/31/2015 | 1.4766 | 1.49352 | 0.01692 | |
| 2/28/2015 | 1.5128 | 1.5299 | 0.0171 | |
| 3/31/2015 | 1.4765 | 1.48092 | 0.00442 | |
| 4/30/2015 | 1.4467 | 1.44905 | 0.00235 | |
| 5/31/2015 | 1.4606 | 1.45959 | -0.00101 | |
| 6/30/2015 | 1.3657 | 1.3894 | 0.0237 | |
| 7/31/2015 | 1.4294 | 1.45721 | 0.02781 | |
| 8/31/2015 | 1.3477 | 1.34289 | -0.00481 | |
| 9/30/2015 | 1.2935 | 1.30644 | 0.01294 | |
| 10/31/2015 | 1.3523 | 1.36275 | 0.01045 | |
| 11/30/2015 | 1.332 | 1.3313 | -0.0007 | |
| 12/31/2015 | 1.2728 | 1.31096 | 0.03816 | |
| 1/31/2016 | 1.2508 | 1.28843 | 0.03763 | |
| 2/29/2016 | 1.2984 | 1.29429 | -0.00411 | |
| 3/31/2016 | 1.3542 | 1.36883 | 0.01463 | |
| 4/30/2016 | 1.3889 | 1.40361 | 0.01471 | |
| 5/31/2016 | 1.3424 | 1.33964 | -0.00276 | |
| 6/30/2016 | 1.3566 | 1.38023 | 0.02363 | |
| 7/31/2016 | 1.3511 | 1.37531 | 0.02421 | |
| 8/31/2016 | 1.4109 | 1.4108 | -1E-04 | |
| 9/30/2016 | 1.3216 | 1.35451 | 0.03291 | |
| 10/31/2016 | 1.3192 | 1.35292 | 0.03372 | |
| 11/30/2016 | 1.4444 | 1.44192 | -0.00248 | |
| 12/31/2016 | 1.4239 | 1.49291 | 0.06901 | |
| 1/31/2017 | 1.4344 | 1.50354 | 0.06914 | |
| 2/28/2017 | 1.4989 | 1.49408 | -0.00482 | |
| 3/31/2017 | 1.4035 | 1.4529 | 0.0494 | |
| 4/30/2017 | 1.3914 | 1.4401 | 0.0487 | |
| 5/31/2017 | 1.3936 | 1.39226 | -0.00134 | |
| 6/30/2017 | 1.3913 | 1.42673 | 0.03543 | |
| 7/31/2017 | 1.4365 | 1.47392 | 0.03742 | |
| 8/31/2017 | 1.4857 | 1.48643 | 0.00073 | |
| 9/30/2017 | 1.4664 | 1.50424 | 0.03784 | |
| 10/31/2017 | 1.4961 | 1.53401 | 0.03791 | |
| 11/30/2017 | 1.5495 | 1.54468 | -0.00482 | |
| 12/31/2017 | 1.4047 | 1.58643 | 0.18173 | |
| 1/31/2018 | 1.5225 | 1.71577 | 0.19327 | |
| 2/28/2018 | 1.4667 | 1.46788 | 0.00118 | |
| 3/31/2018 | 1.4166 | 1.41322 | -0.00338 | |
| 4/30/2018 | 1.3761 | 1.37253 | -0.00357 | |
| 5/31/2018 | 1.3602 | 1.36044 | 0.00024 | |
| 6/30/2018 | 1.2905 | 1.32579 | 0.03529 | |
| 7/31/2018 | 1.3734 | 1.40549 | 0.03209 | |
| 8/31/2018 | 1.4439 | 1.43829 | -0.00561 | |
| 9/30/2018 | 1.4044 | 1.47291 | 0.06851 | |
| 10/31/2018 | 1.3482 | 1.41166 | 0.06346 | |
| 11/30/2018 | 1.4282 | 1.43081 | 0.00261 | |
| 12/31/2018 | 1.4413 | 1.33802 | -0.10328 | |
| 1/31/2019 | 1.4582 | 1.34693 | -0.11127 | |
| 2/28/2019 | 1.4215 | 1.42096 | -0.00054 | |
| 3/31/2019 | 1.3403 | 1.41328 | 0.07298 | |
| 4/30/2019 | 1.4461 | 1.52442 | 0.07832 | |
| 5/31/2019 | 1.3146 | 1.31278 | -0.00182 | |
| 6/30/2019 | 1.3642 | 1.41644 | 0.05224 | |
| 7/31/2019 | 1.3182 | 1.36501 | 0.04681 | |
| 8/31/2019 | 1.2995 | 1.3033 | 0.0038 | |
| 9/30/2019 | 1.282 | 1.33095 | 0.04895 | |
| 10/31/2019 | 1.3106 | 1.35871 | 0.04811 | |
| 11/30/2019 | 1.3548 | 1.35471 | -9E-05 | |
| 12/31/2019 | 1.3034 | 1.3885 | 0.0851 | |
| 1/31/2020 | 1.2907 | 1.37581 | 0.08511 | |
| 2/29/2020 | 1.1829 | 1.18356 | 0.00066 | |
| 3/31/2020 | 1.1947 | 1.04589 | -0.14881 | |
| 4/30/2020 | 1.2299 | 1.0718 | -0.1581 | |
| 5/31/2020 | 1.2142 | 1.21369 | -0.00051 | |
| 6/30/2020 | 1.1016 | 1.15349 | 0.05189 | |
| 7/31/2020 | 1.209 | 1.25837 | 0.04937 | |
| 8/31/2020 | 1.3242 | 1.32334 | -0.00086 | |
| 9/30/2020 | 1.2259 | 1.27493 | 0.04903 | |
| 10/31/2020 | 1.1609 | 1.20307 | 0.04217 | |
| 11/30/2020 | 1.2934 | 1.29285 | -0.00055 | |
| 12/31/2020 | 1.2268 | 1.29348 | 0.06668 | |
| 1/31/2021 | 1.209 | 1.27117 | 0.06217 | |
| 2/28/2021 | 1.2781 | 1.33084 | 0.05274 | |
| 3/31/2021 | 1.3136 | 1.31924 | 0.00564 | |
| 4/30/2021 | 1.41 | 1.41678 | 0.00678 | |
| 5/31/2021 | 1.478 | 1.47538 | -0.00262 | |
| 6/30/2021 | 1.3515 | 1.40629 | 0.05479 | |
| 7/31/2021 | 1.3529 | 1.40427 | 0.05137 | |
| 8/31/2021 | 1.3749 | 1.37328 | -0.00162 | |
| 9/30/2021 | 1.3085 | 1.29948 | -0.00902 | |
| 10/31/2021 | 1.3763 | 1.36265 | -0.01365 | |
| 11/30/2021 | 1.3103 | 1.3079 | -0.0024 | |
| 12/31/2021 | 1.3219 | 1.40244 | 0.08054 | |
| 1/31/2022 | 1.3814 | 1.4682 | 0.0868 | |
| 2/28/2022 | 1.3988 | 1.40608 | 0.00728 | |
| 3/31/2022 | 1.5373 | 1.53819 | 0.00089 | |
| 4/30/2022 | 1.4068 | 1.40708 | 0.00028 | |
| 5/31/2022 | 1.3721 | 1.37196 | -0.00014 | |
| 6/30/2022 | 1.3052 | 1.18349 | -0.12171 | |
| 7/31/2022 | 1.439 | 1.30205 | -0.13695 | |
| 8/31/2022 | 1.3401 | 1.33999 | -0.00011 | |
| 9/30/2022 | 1.282 | 1.27424 | -0.00776 | |
| 10/11/2022 | 1.2772 | 1.27562 | -0.00158 |
| Most Egregious Deltas | ||||
|---|---|---|---|---|
| Date | Tedthedog | cwags02 | Delta | |
| 40908 | 1.1478 | 1.18116 | 0.03336 | |
| 40999 | 1.1427 | 1.21924 | 0.07654 | |
| 41182 | 1.186 | 1.23245 | 0.04645 | |
| 41364 | 1.2966 | 1.36886 | 0.07226 | |
| 41394 | 1.3209 | 1.39678 | 0.07588 | |
| 41455 | 1.3689 | 1.39319 | 0.02429 | |
| 41486 | 1.4144 | 1.44239 | 0.02799 | |
| 41547 | 1.3437 | 1.38547 | 0.04177 | |
| 41578 | 1.3633 | 1.40478 | 0.04148 | |
| 41639 | 1.3178 | 1.403 | 0.0852 | |
| 41670 | 1.2484 | 1.32064 | 0.07224 | |
| 41698 | 1.2868 | 1.37015 | 0.08335 | |
| 41729 | 1.3536 | 1.38911 | 0.03551 | |
| 41759 | 1.396 | 1.4323 | 0.0363 | |
| 41820 | 1.3342 | 1.36979 | 0.03559 | |
| 41851 | 1.322 | 1.35758 | 0.03558 | |
| 41912 | 1.4321 | 1.4547 | 0.0226 | |
| 41943 | 1.4533 | 1.47599 | 0.02269 | |
| 42004 | 1.5433 | 1.55828 | 0.01498 | |
| 42035 | 1.4766 | 1.49352 | 0.01692 | |
| 42063 | 1.5128 | 1.5299 | 0.0171 | |
| 42185 | 1.3657 | 1.3894 | 0.0237 | |
| 42216 | 1.4294 | 1.45721 | 0.02781 | |
| 42277 | 1.2935 | 1.30644 | 0.01294 | |
| 42308 | 1.3523 | 1.36275 | 0.01045 | |
| 42369 | 1.2728 | 1.31096 | 0.03816 | |
| 42400 | 1.2508 | 1.28843 | 0.03763 | |
| 42460 | 1.3542 | 1.36883 | 0.01463 | |
| 42490 | 1.3889 | 1.40361 | 0.01471 | |
| 42551 | 1.3566 | 1.38023 | 0.02363 | |
| 42582 | 1.3511 | 1.37531 | 0.02421 | |
| 42643 | 1.3216 | 1.35451 | 0.03291 | |
| 42674 | 1.3192 | 1.35292 | 0.03372 | |
| 42735 | 1.4239 | 1.49291 | 0.06901 | |
| 42766 | 1.4344 | 1.50354 | 0.06914 | |
| 42825 | 1.4035 | 1.4529 | 0.0494 | |
| 42855 | 1.3914 | 1.4401 | 0.0487 | |
| 42916 | 1.3913 | 1.42673 | 0.03543 | |
| 42947 | 1.4365 | 1.47392 | 0.03742 | |
| 43008 | 1.4664 | 1.50424 | 0.03784 | |
| 43039 | 1.4961 | 1.53401 | 0.03791 | |
| 43100 | 1.4047 | 1.58643 | 0.18173 | |
| 43131 | 1.5225 | 1.71577 | 0.19327 | |
| 43281 | 1.2905 | 1.32579 | 0.03529 | |
| 43312 | 1.3734 | 1.40549 | 0.03209 | |
| 43373 | 1.4044 | 1.47291 | 0.06851 | |
| 43404 | 1.3482 | 1.41166 | 0.06346 | |
| 43465 | 1.4413 | 1.33802 | -0.10328 | |
| 43496 | 1.4582 | 1.34693 | -0.11127 | |
| 43555 | 1.3403 | 1.41328 | 0.07298 | |
| 43585 | 1.4461 | 1.52442 | 0.07832 | |
| 43646 | 1.3642 | 1.41644 | 0.05224 | |
| 43677 | 1.3182 | 1.36501 | 0.04681 | |
| 43738 | 1.282 | 1.33095 | 0.04895 | |
| 43769 | 1.3106 | 1.35871 | 0.04811 | |
| 43830 | 1.3034 | 1.3885 | 0.0851 | |
| 43861 | 1.2907 | 1.37581 | 0.08511 | |
| 43921 | 1.1947 | 1.04589 | -0.14881 | |
| 43951 | 1.2299 | 1.0718 | -0.1581 | |
| 44012 | 1.1016 | 1.15349 | 0.05189 | |
| 44043 | 1.209 | 1.25837 | 0.04937 | |
| 44104 | 1.2259 | 1.27493 | 0.04903 | |
| 44135 | 1.1609 | 1.20307 | 0.04217 | |
| 44196 | 1.2268 | 1.29348 | 0.06668 | |
| 44227 | 1.209 | 1.27117 | 0.06217 | |
| 44255 | 1.2781 | 1.33084 | 0.05274 | |
| 44377 | 1.3515 | 1.40629 | 0.05479 | |
| 44408 | 1.3529 | 1.40427 | 0.05137 | |
| 44561 | 1.3219 | 1.40244 | 0.08054 | |
| 44592 | 1.3814 | 1.4682 | 0.0868 | |
| 44742 | 1.3052 | 1.18349 | -0.12171 | |
| 44773 | 1.439 | 1.30205 | -0.13695 |
Very nice!
Just to be sure I follow:
Smoothing is by taking average of price over 1 to 2 years out, so pin the resulting value at 1.5 years out, then calculate the return of that (with inflation adjustment)
?
Thanks so much!
Good question. Off top of head one would guess your data from M* is more reliable than the free data I pulled off the web. OTOH, the site the free data came from claimed to use careful data sources.
Thanks for sharing those! First time I’ve read that thread.
I suppose that depends on the degree to which their current problems are transient. I can’t contribute an educated viewpoint on that. Some thoughts on this thread:
After Jim posted on 08/02/2020 (copy included in the string of ‘Jim posts’ pasted above) I recalled doing a scatterplot of his binned PriceToBook values and their forward return. It was pretty nice so I recreated it below:
R= -0.943, R^2= 0.8892
Technical Details:
Consider first the large subset of his data where each bin contain 5% of the points along with the associated returns. I’ll call these ‘big bins’. One can simply plot the middle of the big bins versus associated forward return. But Jim also gave five ‘litle bins’, each containing 1% of data at high end, and similarly five little bins at low end. Because the size of these little bins is different, you can’t just throw them in the plot and do the regression line. Weighted least square would work for the regression, but for a simple post let’s keep it simple. So, for the middle of the small bins, take their median and also take the median of their associated returns. This adds two more datapoints to the above table, one at low tail that has 5% of data and similarly one at high tail that has 5% of data. The plot and regression line, above, is of this data, here it is:
PBV_Bin Return
1.1220 24.2
1.1740 21.5
1.2210 17.4
1.2760 11.6
1.3215 7.7
1.3500 12.2
1.3705 10.9
1.3900 7.5
1.4160 8.3
1.4440 9.0
1.4660 8.6
1.4875 9.8
1.5120 8.9
1.5380 3.6
1.5775 0.8
1.6250 1.9
1.6675 0.5
1.7150 -0.9
1.7730 -1.5
Comment 1:
In Jim’s posts that were copied earlier in this thread (above), someone asked if Jim’s PBV value agreed with the PBV value that they got from ycharts.com The reply was to the effect of “similar”, and a comment/question about what date gets assigned to a book value, is it the date that the value was actually published and hence useable or did a data vendor assign the book value to the date for the end of that quarter? Given that price can wiggle over a short time, a difference in dates could create differences in P/BV. In more examples of different data sources and somewhat different data: I pulled down some P/BV data from a site that claimed to use reputable data sources while cwags02 pulled data down from reputable Morningstar. These two data sets are also similar but with differences.
Comment 2:
Financial data is both voluminous and messy, so there’re lots of numbers which at first blush seems nice, but in fact they are dirty numbers that include subtlties related to recording (also outright errors), as well as outliers due to the irrationality of the market. Real trends can be hard to discern and it’s easy to get misled by outliers or other quirks.
Being very certain about a general qualitative conclusion e.g. “this bin of low present peakPtoBV for Berkshire has been associated with future high returns” seems better than being uncertain about a very attractive quantitative conclusion.
To this end, smoothing, binning and other filtering of the data seems appropriate.
This is super cool!
So eyeballing your chart it looks like forward returns are expected to be low 20%s in the next 1.5 years if history repeats?
And it’s down even more. Tough time to have the ticker KMX.
Jim’s original post, that I had earlier pasted above with a bunch of his other posts, is pasted again below for convenience, then some comments:
Author: mungofitch Date: 08/02/2020 11:50 AM Msg: BH-255100
Subject: Observations Recs: 40
No, not observations in the sense of insights, just some numerical observations in the sense of a science experiment.
I was doing some fiddling about and thought others might find these numbers interesting.
Given the ratio of market price to “peak to date” book per share on any given date in the past, look at the average market price 1-2 years later.
So, it’s a look ahead period of average 1.5 years.
The stock prices and resulting returns are adjusted for inflation and annualized.
Starting dates 2002-01-02 through 2018-07-31
Initial ratio of price to peak book Avg 1.5yr return
1% of time P/peakB from 0.928 to 1.076 30.3%/year
1% of time P/peakB from 1.076 to 1.112 24.2%
1% of time P/peakB from 1.112 to 1.132 24.9% We are in this bucket right now
1% of time P/peakB from 1.132 to 1.144 24.0%
1% of time P/peakB from 1.144 to 1.154 23.5%
5% of time P/peakB from 1.154 to 1.194 21.5%
5% of time P/peakB from 1.194 to 1.248 17.4%
5% of time P/peakB from 1.248 to 1.304 11.6%
5% of time P/peakB from 1.304 to 1.339 7.7%
5% of time P/peakB from 1.339 to 1.361 12.2%
5% of time P/peakB from 1.361 to 1.380 10.9%
5% of time P/peakB from 1.380 to 1.400 7.5%
5% of time P/peakB from 1.400 to 1.432 8.3%
5% of time P/peakB from 1.432 to 1.456 9.0%
5% of time P/peakB from 1.456 to 1.476 8.6%
5% of time P/peakB from 1.476 to 1.499 9.8%
5% of time P/peakB from 1.499 to 1.525 8.9%
5% of time P/peakB from 1.525 to 1.551 3.6%
5% of time P/peakB from 1.551 to 1.604 0.8%
5% of time P/peakB from 1.604 to 1.646 1.9%
5% of time P/peakB from 1.646 to 1.689 0.5%
5% of time P/peakB from 1.689 to 1.741 -0.9%
5% of time P/peakB from 1.741 to 1.805 -1.5%
1% of time P/peakB from 1.805 to 1.818 2.3%
1% of time P/peakB from 1.818 to 1.830 -2.8%
1% of time P/peakB from 1.830 to 1.846 -0.3%
1% of time P/peakB from 1.846 to 1.882 -6.9%
1% of time P/peakB from 1.882 to 2.032 -8.0%
I used the “peak to date” book value for the simple reason that dips in book per share have always been transient in the past and using the peak-to-date value gives a stronger predictor of forward returns.
Although I’d like to know precisely how he arrived at those numbers (I give my interpretation of his process below), I suspect he’d be the first person to say that it’s just one approach, and there could be a variety of ways to analyze this i.e. no one “right way”, however some ways may be more informative than others. So even if he did something different in detail from how I interpreted it below, just picking up on his idea of using an average forward return and of binning and going from there, seems useful. But we can always ask him about details of his analysis on his site.
Anyway, I interpreted Jim’s post to mean the following:
The values in his column “Avg 1.5yr return” are from taking the average of prices from 1 yr out to 2 yrs out from the date of a given P/peakB, then computing a return from that average price where the date of this (average) price is pegged at 1.5 years out, then annualize this return.
Each P/peakB value therefore gets associated with it an annualized return from “average market price 1-2 years later”.
He then makes bins of P/peakB values. Most of the are bins contain 5% of the data, let’s pick his first 5% bin from the above table as an example, here it is:
“5% of time P/peakB from 1.154 to 1.194 21.5%”
This example bin contains 5% of the data, i.e. a bunch of pairs (P/peakB,Average1.5YrReturn) where each P/peakB of the pair falls between 1.154 to 1.194
The reason for the funny range numbers, (1.154,1.194), is that these ranges are adjusted so that each bin contain 5% of the data (ordered on P/peakB). We’re familiar with this from ‘quartiles’, where ranges are given that contain the first 0-25%, 25%-50%, 50%-75%, and 75%-100% of ordered data.
He matches this bin with a return of 21.5%
Where’s 21.5% come from? I’m assuming that “21.5” is the average of the “1.5 Yr” returns associated with each P/peakB in the bin.
BTW, one could replace all instances of ‘average’ with ‘median’.
Continuing, look at the first 5% bin, above. The average of 1.154 and 1.194 is 1.174
That’s why you see 1.174 paired with 21.5 in the table that I gave above and that I scatterplotted.
Because the 1% bins in the low tail and in the high tail contain less data than 5% bins, I used a method to group together these smaller ‘tail bins’ to each contain 5% of the data. This method results in two more bins: one of which is for all the low tail points and the other for all the high tail, and all bins now contain the same amount of data (including the tail bins), so can plot and do the regression line that I gave above.
Comment:
Jim’s general style of doing this analysis seems to be an example of one way to get a handle on messy financial data: first, carefully quantitate a qualitative conclusion.
After you have confidence in your qualitative understanding of this messy data, you can feel more comfortable about more specific quantitative approaches. If you jump into messy financial data without a good qualitative idea of what’s going on, there’s a danger that it can be easy to be misled among the mess.
I would do exactly what you describe.
@johnIII Bueno!
@cwags02
Regarding Morningstar data, here’s some fun: 
I have a friend with a paid subscription and he pulled down some BRK-A data for me with PriceToBook starting 04/30/1990.
Here’s a snippet from his M* data for the year 2011:
Date,Name,Close,High,Low,Open,Volume,PB
01/31/2011,Berkshire Hathaway Inc Class A,“122,425”,“125,036”,“118,792”,“120,550”,“12,735”,1.282568,
02/28/2011,Berkshire Hathaway Inc Class A,“131,300”,“131,463”,“123,000”,“123,000”,“9,847”,1.375875,
03/31/2011,Berkshire Hathaway Inc Class A,“125,300”,“131,400”,“121,500”,“131,400”,“10,981”,1.290677,
04/29/2011,Berkshire Hathaway Inc Class A,“124,750”,“126,100”,“119,683”,“126,000”,“10,616”,
05/31/2011,Berkshire Hathaway Inc Class A,“118,775”,“123,830”,“115,860.01”,“123,100”,“8,265”,1.223571,
06/30/2011,Berkshire Hathaway Inc Class A,“116,105”,“118,340”,“109,925”,“118,020”,“11,233”,1.176152,
07/29/2011,Berkshire Hathaway Inc Class A,“111,500”,“117,250”,“111,246”,“115,230”,“9,086”,
08/31/2011,Berkshire Hathaway Inc Class A,“109,769”,“113,445”,“100,265”,“113,400”,“19,962”,1.111976,
09/30/2011,Berkshire Hathaway Inc Class A,“106,800”,“111,751”,“98,952”,“109,374”,“15,839”,1.102442,
10/31/2011,Berkshire Hathaway Inc Class A,“116,950”,“120,755”,“104,701”,“107,600”,“14,100”,1.206926,
11/30/2011,Berkshire Hathaway Inc Class A,“118,500”,“118,500”,“110,092”,“114,329.99”,“13,586”,1.222922,
12/30/2011,Berkshire Hathaway Inc Class A,“114,755”,“118,373”,“110,741”,“116,900”,“10,479”,
Interestingly, this Morningstar data doesn’t agree in detail with your Morningstar data, although very similar.
One obvious difference is that the data my friend pulled down has a gap for “04/29/2011” (the last trading day of the month) whereas your Morningstar data has a value for 04/30/2011 (the last calendar day of that month). Friend’s data also has a gap for “07/29/2011” (last trading day of month) whereas your data has a value for 7/31/2011 (last calendar day). Friends data has a gap for “12/30/2011” (last trading day of that month) whereas your Morningstar has a value for “12/31/2011” (last calendar day of month).
I read it into “R”, despite the gaps, friend’s Morningstar data does have values for all end-of-quarter dates.
These seem to be different versions of Morningstar data i.e. slightly different values, different choice of end of month reporting dates, etc.
Why friend’s Morningstar version has gaps for some dates is unclear. Also unclear is why they apparently randomly pick some dates to fill with values but not others (this latter point is more clear in other years where they sometimes have the end-of-quarter value alone, sometimes with a month or two around it, sometimes with five months.
Argh.
Here’s a link to an old KMX thread. It appears Jim entered KMX pretty heavily at prices over $90.
VL has slashed their earnings estimates to $4.55/sh for 2022 and $4.90/sh for 2023. If correct, that would indicate this dip is not transient. They give it a timeliness score of 1, FWIW.
I’m starting to fiddle with some of this analysis myself, many thanks for all your excellent work and the inspiration!
Question:
Your plot of inflation adjusted BRK forward 1year return versus (peak) bookvalue has a box in lower right hand side that references regression of return against basically log(book). I’m assuming the scatterplot also uses log(book) but there’s a typo in the x-axis label that left out “log”?
To cut to the chase, I’m not sure ‘return’ should be regressed or plotted against ‘log(book)’, rather it should simply be ‘book’. I think @johnIII in an earlier comment was introducing a log, but that was in a different context. Maybe he wants to comment?
And just so I’m clear, do the KMX plots use log() at all?
I would agree. I might use log(y) for a time series y that I expected to grow exponentially (e.g. price) but wanted to fit a line to.
FYI, I sent following to Jim (mungofitch):
Jim,
You may find this thread initiated by cwags02, of interest. He’s trying to reproduce your analysis of forward BRK return in relation to book value, and then apply it to KMX using ‘sales’ instead of ‘book’. Given my interest in your original post, I’m kibbitzing a bit and slowly starting to do some analysis as well:
#1Year Berkshire Price Prediction - Inflation +17% From 10/22 - #59 by tedthedog
I understand that others have contacted you about ‘lemonfool’ as a possible replacement to discuss BRK and perhaps other matters.
Cheers!
I won’t copy his reply because I don’t forward/copy other people’s emails w/o asking, but it was to the effect that this month he’s very busy with real world issues.