WEBVTT 00:01.240 --> 00:06.100 Welcome to the final conclusion covering actually the last three sections. 00:06.400 --> 00:11.800 And in the last couple of videos we have seen that there are severe pitfalls when forecasting future 00:11.800 --> 00:19.630 portfolios and what we did in the last videos is also quite a mean variance optimization and actually 00:19.630 --> 00:26.530 Modern Portfolio Theory asset pricing models and the mean variance optimization are one of the most 00:26.620 --> 00:33.990 misunderstood concepts in finance because typically they are explained and derive with the forecasts 00:33.990 --> 00:38.120 or expectations in a forward looking scenario. 00:38.320 --> 00:44.530 And this is done because typically you require research for model assumptions to derive these models 00:45.280 --> 00:49.440 and these assumptions only make sense and forward looking scenarios. 00:49.450 --> 00:56.730 However we have seen that explaining these theories and models with past performance works as well and 00:56.720 --> 01:02.350 it's even more intuitive as we didn't need any of those unrealistic assumptions. 01:02.350 --> 01:08.760 So we simply analyzed past data and tried to find some patterns in the past data. 01:08.770 --> 01:12.270 And this worked pretty good actually. 01:12.280 --> 01:17.740 Again these theories and models are typically explained in a forward looking scenario. 01:17.740 --> 01:24.850 However it is actually a rarely used for real world asset allocation problems because there are severe 01:24.850 --> 01:26.050 pitfalls. 01:26.050 --> 01:29.750 Well at least they are not used in their simplest form. 01:29.830 --> 01:34.020 So let's have a look now at the pitfalls and we have pitfalls. 01:34.030 --> 01:34.970 Number one. 01:35.110 --> 01:38.690 So forecasting the future is always trickier. 01:38.770 --> 01:44.970 And actually the outcomes of the mean variance optimization are highly sensitive to the inputs. 01:45.490 --> 01:50.560 So the general rule garbage in garbage out also applies here. 01:50.660 --> 01:58.940 And first of all it's hard if not impossible to forecast the future stock returns then we used the past 01:58.940 --> 02:02.870 variances and covariance also for future predictions. 02:02.870 --> 02:07.520 However the variance of some covariance is are not stable over time. 02:07.820 --> 02:10.010 So this is a simplification. 02:10.310 --> 02:14.740 And to sum it up forecasting the future is no exact science. 02:17.660 --> 02:23.390 And we've also seen that that mean variance optimization in a forward looking scenario typically leads 02:23.390 --> 02:27.650 to highly concentrated positions and portfolios. 02:27.650 --> 02:33.800 So we have a high concentration and only a few stocks or assets and we have zero allocation to many 02:33.800 --> 02:41.410 other stocks or assets so actually the target of the mean variance optimization is to create diversified 02:41.410 --> 02:42.610 portfolios. 02:42.730 --> 02:49.300 However optimal in a mean variance framework does not necessarily mean practical diversification 02:52.220 --> 02:59.510 and we've also seen that the most sensitive factor are actually the forecasts that returns and again 02:59.510 --> 03:06.930 it's hard if not impossible to forecast the stock returns and actually small adjustments in our forecasts. 03:06.970 --> 03:15.400 Predictions that lead to large changes in allocations and weights and if we were to react to small adjustments 03:15.460 --> 03:24.010 and rebalance or adjust the allocations then uh the possible efficiency gains by having a more efficient 03:24.010 --> 03:25.090 portfolio. 03:25.090 --> 03:31.300 So of course there's actually no guarantee to get a better portfolio but uh the potential efficiency 03:31.300 --> 03:38.400 gains are typically offset by costs and taxes that become due when we heavily adjust and rebalance. 03:38.440 --> 03:40.680 The allocations in our portfolios. 03:41.470 --> 03:49.240 So excessive trading our turn over in our portfolio is in many cases not a good idea because potential 03:49.270 --> 03:53.890 positive effects are typically offset or more than offset by costs. 03:53.920 --> 03:54.790 And Texas 03:58.850 --> 04:05.420 so let's have a look at portfolios of institutional investors like pension funds insurances and so on. 04:05.420 --> 04:08.320 And typically they are highly constrained. 04:08.420 --> 04:16.520 So there are regulatory or legal constraints and typically they're very restrictive limits on the allocation 04:16.550 --> 04:20.270 or rates of asset classes or assets. 04:20.270 --> 04:27.560 So for example a pension funds are insurances have had limits for stock so for example 15 percent only 04:29.530 --> 04:32.800 so these investors pretty much have predetermined buckets. 04:32.860 --> 04:40.360 Let's say 15 percent stocks their off maximum 5 percent foreign stocks and at least the 60 percent high 04:40.360 --> 04:43.330 quality is some rain and corporate bonds. 04:43.600 --> 04:49.450 And that's a typical workflow when evaluating whether an additional asset class will be added and a 04:49.450 --> 04:55.750 few years ago I advised that German insurance companies and pension funds on whether to add a new asset 04:55.750 --> 05:02.510 class to the portfolio in this case that was aircraft backed mortgage loans and bonds. 05:02.650 --> 05:09.460 And this is typically done with a backward looking portfolio construction or to say a backward looking 05:09.460 --> 05:11.460 mean variance optimization. 05:11.620 --> 05:19.690 And we actually added up to 5 percent of this new asset class to the past the portfolio and we created 05:19.780 --> 05:21.760 the past the efficient frontier. 05:21.760 --> 05:27.820 And in case that the efficient frontier moved significantly to the upper left corner then the asset 05:27.820 --> 05:29.590 class will be added. 05:29.620 --> 05:32.680 So let's have a look at an example here. 05:32.680 --> 05:37.640 We know that graph already from the last video and this is actually forward looking. 05:38.050 --> 05:45.520 But let's assume that the blue portfolios are past portfolios without the new asset class and the red 05:45.520 --> 05:52.970 portfolios are also simulations of the past including the new asset class or maximum 5 percent. 05:52.990 --> 05:59.590 And here we can see that the efficient frontier significantly moves to the upper left corner here. 05:59.890 --> 06:03.180 So this is a significant difference here. 06:03.280 --> 06:07.840 And consequently the new asset class will be added to the portfolio. 06:09.460 --> 06:15.340 So that's a typical application of the mean variance optimization for institutional investors. 06:15.340 --> 06:24.680 And again this is based on past performance and it's not about predicting the future all right and I 06:24.680 --> 06:31.400 said before that the mean variance optimization is uh rarely used in its simplest form. 06:31.550 --> 06:38.120 But uh there are some advanced approaches that can be observed in the market but they are definitely 06:38.120 --> 06:40.230 beyond the scope of this course. 06:40.250 --> 06:42.860 However at least I want to mention them here. 06:42.920 --> 06:45.350 So we have a reverse optimization. 06:45.360 --> 06:48.250 Then we have the black letter man model. 06:48.250 --> 06:54.470 Then we have read sampling or Monte Carlo simulations and also adding constraints to our mean variance 06:54.470 --> 06:55.970 optimization. 06:55.970 --> 07:01.810 And in particular for individual investors that s also the concept of gold based investing. 07:01.910 --> 07:04.210 And actually many more advanced concepts. 07:04.250 --> 07:09.000 So explaining those advanced approaches that would take another 20 hours. 07:09.200 --> 07:15.890 And uh the question is uh whether those advanced methods do guarantee better results. 07:15.890 --> 07:22.980 And uh this is definitely not the case and the final question is here now what should I do with my stock 07:22.980 --> 07:24.510 portfolio. 07:24.510 --> 07:32.370 And at least for an individual investor it might be best to simply trust the market expectations. 07:32.490 --> 07:39.420 So current stock prices that you can observe in the market are formed by the ever it's market expectations 07:39.440 --> 07:41.230 or market consensus. 07:41.370 --> 07:47.460 And if the market believes that the stock price is overpriced or will perform poorly in the near future 07:47.790 --> 07:50.940 then the price would probably fall immediately. 07:50.940 --> 07:57.420 So each stock price reflects the fair price of that stock based on market expectations. 07:57.510 --> 08:04.340 And of course this does not necessarily mean that the market is right but uh it is probably the best 08:04.500 --> 08:06.610 forecast that you can get. 08:06.720 --> 08:12.000 And as a consequence the debates are proportions of each and every stock in your portfolio should be 08:12.000 --> 08:15.470 determined by the proportional market capitalization. 08:15.480 --> 08:21.690 And that means that you should simply hold and fund or an ETF that closely follows a proto market cap 08:21.690 --> 08:22.720 weighted index. 08:23.160 --> 08:30.150 So for example the S&P 500 index and this is actually the optimal forward looking stock portfolio based 08:30.150 --> 08:31.910 on market consensus. 08:31.920 --> 08:39.590 However this does of course not mean that this will be the optimal portfolio from an ex post perspective. 08:39.630 --> 08:41.790 Actually that is pretty unlikely. 08:42.150 --> 08:48.450 But as I said before it's still the best possible forecast you can get unless you know it better. 08:48.750 --> 08:55.720 And finally this is the Perry toe principle which leads me to one of the first videos of this costs. 08:55.800 --> 09:03.070 So typically 85 percent of the portfolio performance or the success of your portfolio are driven by 09:03.070 --> 09:11.560 a few simple rules that can be actually explained in two minutes you might know this comparison already 09:11.560 --> 09:12.570 on the left hand side. 09:12.580 --> 09:20.380 We have an aggregation of the past performance of the S&P 500 index and on the right hand side the same 09:20.380 --> 09:22.700 analysis for the T stock. 09:22.810 --> 09:26.170 And from that comparison we can derive some simple rules. 09:26.650 --> 09:31.990 So holding a diversified portfolio is better than holding a concentrated portfolio. 09:32.230 --> 09:39.100 And the second rule is constructing an appropriate portfolio based on your individual investment horizon. 09:39.190 --> 09:45.640 So with a short horizon of one or two years you should to better reduce the proportion in stocks and 09:45.730 --> 09:49.700 to allocate your wealth to short term high quality bonds. 09:49.960 --> 09:55.840 And on the other hand side if you have a long investment horizon then you will benefit from the time 09:55.840 --> 09:57.600 diversification effect. 09:57.670 --> 10:02.950 And as a consequence you should increase the proportion of stocks. 10:03.040 --> 10:09.550 And finally as a third rule avoid excessive trading and rebalancing as this only results in costs and 10:09.550 --> 10:17.350 Texas so that's my personal view and of course feel free to form your own opinion and make your own 10:17.350 --> 10:18.610 analysis. 10:18.610 --> 10:25.330 You are now equipped with all tools to create analyze and optimize portfolios for past portfolios but 10:25.330 --> 10:28.240 also far forward looking portfolios. 10:28.240 --> 10:33.940 And we are finishing now so thanks for watching the last sections and looking forward to seeing you 10:34.360 --> 10:36.550 in the next section on interactive plots.