WEBVTT 00:00:00.000 --> 00:00:02.259 OK, let's unpack this with a phrase that is, 00:00:02.339 --> 00:00:06.139 I think, older than finance, older than banking, 00:00:06.419 --> 00:00:08.460 probably older than written trade laws itself. 00:00:08.720 --> 00:00:10.160 You're talking about the eggs in the basket. 00:00:10.300 --> 00:00:12.339 Don't put all your eggs in one basket. Exactly. 00:00:12.439 --> 00:00:15.039 It is the ultimate piece of folk wisdom, right? 00:00:15.119 --> 00:00:17.859 It's applied to investing all the time. And yet 00:00:17.859 --> 00:00:21.480 the the underlying mathematics, the economics 00:00:21.480 --> 00:00:24.800 of why that simple principle actually works are, 00:00:24.920 --> 00:00:27.780 well, they're often forgotten. even by really 00:00:27.780 --> 00:00:30.539 sophisticated investors. And that's our mission 00:00:30.539 --> 00:00:32.820 for this deep dive. We want to gain knowledge 00:00:32.820 --> 00:00:36.759 quickly, but also thoroughly about this foundational 00:00:36.759 --> 00:00:39.579 principle, financial diversification. We're going 00:00:39.579 --> 00:00:41.600 to try and go beyond the cliche. Right. We want 00:00:41.600 --> 00:00:43.679 to understand the real mechanisms of risk reduction, 00:00:43.899 --> 00:00:46.219 what the limits of diversification are, and maybe 00:00:46.219 --> 00:00:48.340 most importantly, debunk some of these persistent 00:00:48.340 --> 00:00:51.600 myths about how risk behaves over time. At its 00:00:51.600 --> 00:00:53.969 core, it's just a risk reduction technique. That's 00:00:53.969 --> 00:00:56.829 all it is. It's the structured process of allocating 00:00:56.829 --> 00:00:59.090 your capital across a whole variety of assets 00:00:59.090 --> 00:01:00.789 and exposures. And the goal is to reduce your 00:01:00.789 --> 00:01:03.469 exposure to any single point of failure. To any 00:01:03.469 --> 00:01:06.310 one particular asset failure or, God forbid, 00:01:06.549 --> 00:01:09.349 a catastrophic risk event. And this is where 00:01:09.349 --> 00:01:12.170 the conversation and, frankly, the surprise really 00:01:12.170 --> 00:01:15.909 begins. You might think, just intuitively, that 00:01:15.909 --> 00:01:19.609 if you combine, say, 10 risky things, the total 00:01:19.609 --> 00:01:21.890 risk is just the average risk of those 10 things. 00:01:22.049 --> 00:01:24.150 That's the common sense assumption. But because 00:01:24.150 --> 00:01:26.670 of the way different asset returns interact with 00:01:26.670 --> 00:01:30.609 each other, a properly diversified portfolio 00:01:30.609 --> 00:01:34.180 exhibits this remarkable property. it will have 00:01:34.180 --> 00:01:37.439 less variance, so less volatility, less measured 00:01:37.439 --> 00:01:40.640 risk, than the weighted average variance of its 00:01:40.640 --> 00:01:43.159 parts. It's so counterintuitive, isn't it? It 00:01:43.159 --> 00:01:45.799 is. You are actively combining volatile assets, 00:01:45.859 --> 00:01:47.980 but the combination is somehow calmer than the 00:01:47.980 --> 00:01:50.519 sum of its parts. Often, and this is the part 00:01:50.519 --> 00:01:52.840 that really gets people, the volatility of your 00:01:52.840 --> 00:01:55.400 entire diversified basket can actually be less 00:01:55.400 --> 00:01:57.500 than the volatility of the least volatile asset 00:01:57.500 --> 00:01:59.420 you hold inside it. So long as they don't move 00:01:59.420 --> 00:02:02.019 in perfect lockstep. Exactly. Provided they don't 00:02:02.019 --> 00:02:04.459 move in lockstep. This lower variance, that's 00:02:04.459 --> 00:02:06.579 the prize we're all chasing. And we should probably 00:02:06.579 --> 00:02:09.319 clarify for those of you following along that 00:02:09.319 --> 00:02:11.620 diversification isn't the only way to manage 00:02:11.620 --> 00:02:14.439 risk, even if it is the most sort of foundational. 00:02:15.000 --> 00:02:18.120 The other main technique is hedging. Right. Hedging 00:02:18.120 --> 00:02:20.979 is much more of a tactical maneuver. It's about 00:02:20.979 --> 00:02:24.139 taking an offsetting position, maybe using a 00:02:24.139 --> 00:02:26.479 financial instrument like a derivative. To mitigate 00:02:26.479 --> 00:02:30.300 a specific. A known risk factor, exactly, like 00:02:30.300 --> 00:02:32.719 currency risk or interest rate movements for 00:02:32.719 --> 00:02:35.580 a very specific period. Diversification, on the 00:02:35.580 --> 00:02:38.180 other hand, is strategic. It's a structural approach 00:02:38.180 --> 00:02:41.659 to just diluting risk across multiple, hopefully 00:02:41.659 --> 00:02:44.379 non -correlated assets, making the whole portfolio 00:02:44.379 --> 00:02:47.240 inherently more robust. Rather than just shielding 00:02:47.240 --> 00:02:49.919 it from one specific danger. Precisely. Okay, 00:02:49.979 --> 00:02:51.719 so let's use the simplest and maybe the most 00:02:51.719 --> 00:02:54.439 terrifying example to illustrate this need for 00:02:54.439 --> 00:02:57.740 structural diversification. Imagine you the listener 00:02:57.740 --> 00:03:01.419 hold an entirely undiversified portfolio. You 00:03:01.419 --> 00:03:03.460 decide to put your entire retirement savings 00:03:03.460 --> 00:03:06.539 into just one company stock. You're facing maximum 00:03:06.539 --> 00:03:09.400 idiosyncratic risk. And we aren't talking about, 00:03:09.400 --> 00:03:12.020 you know, hypothetical disasters here. It is 00:03:12.020 --> 00:03:15.000 historically common for a single, even a large 00:03:15.000 --> 00:03:19.159 cap stock, to drop 50 % or more in a single year. 00:03:19.360 --> 00:03:22.280 Oh, easily. A specific scandal, a major product 00:03:22.280 --> 00:03:25.560 recall, some new regulation, or just terrible 00:03:25.560 --> 00:03:27.740 management. It happens all the time. And as you've 00:03:27.740 --> 00:03:30.680 pointed out before, recovering from a 50 % drop 00:03:30.680 --> 00:03:34.080 requires a 100 % gain just to get back to even. 00:03:34.159 --> 00:03:37.280 The path back is... brutally difficult. So now 00:03:37.280 --> 00:03:39.759 contrast that nightmare scenario with one where 00:03:39.759 --> 00:03:44.400 you hold a portfolio of, say, 20 stocks randomly 00:03:44.400 --> 00:03:46.680 selected from different industries, company sizes, 00:03:46.780 --> 00:03:48.979 maybe different countries. The probability of 00:03:48.979 --> 00:03:52.060 that entire aggregated basket dropping 50 percent 00:03:52.060 --> 00:03:54.659 in the same year all at once due to unrelated 00:03:54.659 --> 00:03:57.979 catastrophic failures. It just becomes astronomically 00:03:57.979 --> 00:03:59.939 small. The diversification kind of smooths everything 00:03:59.939 --> 00:04:02.039 out. It eliminates the trends that are specific 00:04:02.039 --> 00:04:05.060 to one industry or one company class. That is 00:04:05.060 --> 00:04:07.719 the immediate visceral benefit you get. You're 00:04:07.719 --> 00:04:09.819 no longer betting on the fate of a single CEO. 00:04:10.060 --> 00:04:12.330 You're betting on the long term health. of the 00:04:12.330 --> 00:04:15.550 entire economy. And this structural dilution 00:04:15.550 --> 00:04:18.290 of risk, it's not just limited to stocks and 00:04:18.290 --> 00:04:21.610 bonds in one country. Since, what, the mid -70s, 00:04:21.610 --> 00:04:24.209 the concept has really pushed aggressively toward 00:04:24.209 --> 00:04:26.910 geographic diversification. That was a pivotal 00:04:26.910 --> 00:04:28.829 shift, especially for institutional investment. 00:04:29.230 --> 00:04:32.430 Managers began to heavily advocate for investing 00:04:32.430 --> 00:04:35.209 in emerging markets. You know, think Asia, Latin 00:04:35.209 --> 00:04:37.550 America, parts of Eastern Europe. Places where 00:04:37.550 --> 00:04:39.889 the underlying economic growth rates were often 00:04:39.889 --> 00:04:42.639 much higher. Higher than in the established G7 00:04:42.639 --> 00:04:45.259 nations exactly. And the argument was simple. 00:04:45.639 --> 00:04:48.259 By capturing those higher return rates while 00:04:48.259 --> 00:04:50.839 at the same time reducing your overall portfolio 00:04:50.839 --> 00:04:53.839 risk because those foreign economies don't move 00:04:53.839 --> 00:04:56.360 in perfect lockstep with your home market. You 00:04:56.360 --> 00:04:59.420 generate superior risk adjusted returns. That's 00:04:59.420 --> 00:05:01.579 the holy grail. So if the U .S. market is struggling 00:05:01.579 --> 00:05:03.779 with, I don't know, some internal political crisis, 00:05:03.959 --> 00:05:07.079 your exposure to booming manufacturing in Southeast 00:05:07.079 --> 00:05:10.170 Asia helps cushion that blow. You're using the 00:05:10.170 --> 00:05:13.089 global economic cycle to flatten your personal 00:05:13.089 --> 00:05:16.329 volatility curve. You're recognizing that the 00:05:16.329 --> 00:05:18.730 events that influence one market are not necessarily 00:05:18.730 --> 00:05:21.269 the same events that influence another. OK, so 00:05:21.269 --> 00:05:23.509 we've firmly established that diversification 00:05:23.509 --> 00:05:26.949 reduces risk. It reduces volatility. But now 00:05:26.949 --> 00:05:29.709 we have to deal with the inevitable objection 00:05:29.709 --> 00:05:32.790 from any aggressive investor. What does this 00:05:32.790 --> 00:05:35.720 mean for my returns? The big question. If I'm 00:05:35.720 --> 00:05:38.120 deliberately choosing to spread my capital around, 00:05:38.399 --> 00:05:41.939 am I diluting my potential gains? Am I sacrificing 00:05:41.939 --> 00:05:44.300 my chance to find that next hundred bagger stock? 00:05:45.839 --> 00:05:48.399 Let's turn our attention from safety to the performance 00:05:48.399 --> 00:05:51.040 side of spreading your bets. This is truly the 00:05:51.040 --> 00:05:53.519 essential, maybe even existential question for 00:05:53.519 --> 00:05:55.959 investors considering diversification. It speaks 00:05:55.959 --> 00:05:58.480 to a core psychological hurdle we all have. It 00:05:58.480 --> 00:06:00.360 feels like you're choosing to be average. It 00:06:00.360 --> 00:06:02.800 can feel that way. Let's look at the math first. 00:06:03.230 --> 00:06:05.430 When we talk about expected returns, and by that 00:06:05.430 --> 00:06:08.050 I mean our prior rational data -driven expectations 00:06:08.050 --> 00:06:10.949 of the average returns on all assets, if those 00:06:10.949 --> 00:06:13.129 expectations are identical across all the assets 00:06:13.129 --> 00:06:15.769 we choose, then the expected return on the diversified 00:06:15.769 --> 00:06:18.069 portfolio will also be identical to the expected 00:06:18.069 --> 00:06:20.490 return on the undiversified one. Okay, so if 00:06:20.490 --> 00:06:23.389 I buy 20 stocks that I think will have an average 00:06:23.389 --> 00:06:26.829 return, my portfolio's expected return is still 00:06:26.829 --> 00:06:29.620 just that average return. Diversification doesn't 00:06:29.620 --> 00:06:31.639 give me a free lunch where I get less risk and 00:06:31.639 --> 00:06:34.939 more average return. That is the tradeoff. Diversification 00:06:34.939 --> 00:06:37.899 is not an engine for increasing your average 00:06:37.899 --> 00:06:41.360 returns. Its core purpose is purely statistical. 00:06:41.699 --> 00:06:44.779 It narrows the range of possible outcomes. It 00:06:44.779 --> 00:06:47.019 brings the goalposts closer together. Exactly. 00:06:47.040 --> 00:06:50.120 The extremes, both good and bad, are mitigated. 00:06:50.259 --> 00:06:52.480 Okay, let's unpack that mitigation. Why do I 00:06:52.480 --> 00:06:54.939 lose the extremes on the high side? That's the 00:06:54.939 --> 00:06:56.980 part that hurts. Because you are now holding 00:06:56.980 --> 00:06:59.620 a weighted average of everything, in any given 00:06:59.620 --> 00:07:01.819 period, a group of investments will naturally 00:07:01.819 --> 00:07:04.319 have a spread of performance. There's going to 00:07:04.319 --> 00:07:06.639 be a winner, a loser, and everything in between. 00:07:06.860 --> 00:07:08.240 And you don't know which is which ahead of time? 00:07:08.509 --> 00:07:10.470 You don't. So you have to hold the whole basket. 00:07:10.829 --> 00:07:13.389 Therefore, the return on your diversified portfolio 00:07:13.389 --> 00:07:16.290 can never exceed the return of the single top 00:07:16.290 --> 00:07:18.810 performing investment in your basket. Unless 00:07:18.810 --> 00:07:21.110 all the assets return the exact same amount, 00:07:21.250 --> 00:07:24.110 which never happens, the diversified return will 00:07:24.110 --> 00:07:26.490 always be mathematically lower than that single 00:07:26.490 --> 00:07:29.009 highest return. That feels like the hard truth 00:07:29.009 --> 00:07:32.029 of diversification. You are deliberately, consciously 00:07:32.029 --> 00:07:34.949 sacrificing the chance to hit the jackpot. You 00:07:34.949 --> 00:07:37.370 are choosing not to chase the highest possible 00:07:37.370 --> 00:07:40.759 return. And in exchange, you get certainty or 00:07:40.759 --> 00:07:43.360 more certainty. It's the price of safety. It's 00:07:43.360 --> 00:07:46.199 the explicit price of safety. But look at the 00:07:46.199 --> 00:07:48.439 flip side, which is far more critical for long 00:07:48.439 --> 00:07:51.040 term survival. Conversely, the return of your 00:07:51.040 --> 00:07:53.240 diversified portfolio will always be higher than 00:07:53.240 --> 00:07:55.079 that of the single worst performing investment. 00:07:55.300 --> 00:07:57.959 By diversifying, you sacrifice the chance of 00:07:57.959 --> 00:08:00.240 achieving the maximum theoretical gain. But in 00:08:00.240 --> 00:08:02.959 return, you completely avoid the risk of suffering 00:08:02.959 --> 00:08:05.439 the maximum theoretical loss. That distinction 00:08:05.439 --> 00:08:08.029 is paramount, isn't it? especially for capital 00:08:08.029 --> 00:08:10.889 preservation. You're giving up the tail of maximum 00:08:10.889 --> 00:08:14.009 upside, that lottery ticket win, to completely 00:08:14.009 --> 00:08:16.910 chop off the catastrophic tail of maximum downside. 00:08:17.170 --> 00:08:20.069 It stabilizes your returns. Diversification doesn't 00:08:20.069 --> 00:08:22.569 inherently boost or penalize your expected returns, 00:08:22.790 --> 00:08:25.569 but it makes them vastly more predictable. If 00:08:25.569 --> 00:08:27.529 your goal is preserving capital and ensuring 00:08:27.529 --> 00:08:30.069 a smooth journey toward a financial goal, that 00:08:30.069 --> 00:08:32.669 narrow path provided by diversification is really 00:08:32.669 --> 00:08:35.539 the only viable road. It manages the inherent 00:08:35.539 --> 00:08:39.279 uncertainty of, well, of capitalism. It smooths 00:08:39.279 --> 00:08:41.960 the path. It turns a jagged mountain range of 00:08:41.960 --> 00:08:44.259 potential outcomes into a gentle hill. Which 00:08:44.259 --> 00:08:46.419 naturally leads us to the practical side, the 00:08:46.419 --> 00:08:50.240 how -to. If the goal is stability, how much diversification 00:08:50.240 --> 00:08:52.980 is enough? Is there a magic number of assets 00:08:52.980 --> 00:08:54.919 I need to hold before I can say, you know, I'm 00:08:54.919 --> 00:08:58.039 safe? This is a point of, um... extensive debate 00:08:58.039 --> 00:08:59.879 in financial literature, and it's because there 00:08:59.879 --> 00:09:02.159 is no single magic number that just instantly 00:09:02.159 --> 00:09:04.500 transforms a high -risk portfolio into a low 00:09:04.500 --> 00:09:06.480 -risk one. But there are rules of thumb. There 00:09:06.480 --> 00:09:09.639 are. The number 30, 30 stocks, is often cited 00:09:09.639 --> 00:09:12.360 historically, particularly in the context of 00:09:12.360 --> 00:09:14.480 achieving sufficient diversification within a 00:09:14.480 --> 00:09:17.019 single large equity market like the U .S. Okay, 00:09:17.080 --> 00:09:19.299 so why 30? Is that based on some observation, 00:09:19.360 --> 00:09:22.509 or is it just a nice round number? It's an approximation, 00:09:22.789 --> 00:09:25.090 really, informed by empirical data, which we'll 00:09:25.090 --> 00:09:27.389 definitely get into later. But the consensus 00:09:27.389 --> 00:09:29.429 among quantitative researchers suggests that 00:09:29.429 --> 00:09:31.590 the vast majority of the risk reduction benefit 00:09:31.590 --> 00:09:35.429 actually arrives much earlier than 30 stocks. 00:09:35.549 --> 00:09:38.309 How much earlier? A key finding established as 00:09:38.309 --> 00:09:41.830 early as 1985 and validated many times since 00:09:41.830 --> 00:09:44.210 showed that most of the measurable value you 00:09:44.210 --> 00:09:46.730 get from diversification. So the elimination 00:09:46.730 --> 00:09:50.090 of that. company specific risk is achieved by 00:09:50.090 --> 00:09:52.529 incorporating just the first 15 or 20 different 00:09:52.529 --> 00:09:54.950 stocks into a portfolio. So the heavy lifting 00:09:54.950 --> 00:09:57.769 is done by the 20th stock and the 30th stock 00:09:57.769 --> 00:09:59.769 is just kind of polished. That's a perfect way 00:09:59.769 --> 00:10:02.490 to put it. The overall principle holds. Adding 00:10:02.490 --> 00:10:05.230 more stocks generally leads to lower price volatility. 00:10:05.470 --> 00:10:07.789 But the marginal benefit, the risk reduction 00:10:07.789 --> 00:10:10.570 you gain by adding the 31st stock versus the 00:10:10.570 --> 00:10:14.110 16th, it's statistically negligible. That efficiency 00:10:14.110 --> 00:10:17.389 is so important for the average investor. If 00:10:17.389 --> 00:10:20.129 I only have the resources or, frankly, the time 00:10:20.129 --> 00:10:22.690 to research 20 companies, I've already done most 00:10:22.690 --> 00:10:25.590 of the risk mitigation I need to do. You've neutralized 00:10:25.590 --> 00:10:29.230 what we call security -specific risk. By holding 00:10:29.230 --> 00:10:31.629 20 different companies across different sectors, 00:10:31.809 --> 00:10:35.149 the risk that one specific CEO engages in accounting 00:10:35.149 --> 00:10:37.669 fraud. Or that one specific factory catches fire. 00:10:37.830 --> 00:10:39.629 And shuts down the company's production line. 00:10:39.750 --> 00:10:42.690 That has almost no measurable effect on your 00:10:42.690 --> 00:10:45.789 overall financial well -being. That kind of company 00:10:45.789 --> 00:10:48.570 -specific disaster becomes an isolated, small 00:10:48.570 --> 00:10:51.870 event in your diversified landscape, rather than 00:10:51.870 --> 00:10:54.210 a catastrophe for your entire wealth. So that 00:10:54.210 --> 00:10:57.350 baseline, maybe 15 to 20 assets, becomes the 00:10:57.350 --> 00:10:59.500 critical cycle. psychological threshold where 00:10:59.500 --> 00:11:01.679 you shift from a high risk concentrated approach 00:11:01.679 --> 00:11:05.340 to a structurally sound low idiosyncratic risk 00:11:05.340 --> 00:11:07.879 portfolio. So we've established that, okay, 15 00:11:07.879 --> 00:11:10.340 to 20 assets gives you enough basic diversification 00:11:10.340 --> 00:11:13.100 to handle that company specific risk. But what 00:11:13.100 --> 00:11:14.860 if we want to go further? What if we want to 00:11:14.860 --> 00:11:17.399 achieve maximum diversification? What's the theoretical 00:11:17.399 --> 00:11:19.899 peak of spreading your capital? The theoretical 00:11:19.899 --> 00:11:22.580 ideal, the thing that quants and academics are 00:11:22.580 --> 00:11:25.600 always chasing, is often described as buying 00:11:25.600 --> 00:11:28.519 the market portfolio. That sounds like a concept 00:11:28.519 --> 00:11:30.899 ripped straight from an economics textbook. It 00:11:30.899 --> 00:11:33.720 is, pretty much. The earliest and most influential 00:11:33.720 --> 00:11:36.659 formal definition of this maximum diversification 00:11:36.659 --> 00:11:40.340 came with the advent of the Capital Asset Pricing 00:11:40.340 --> 00:11:43.460 Model, or CAPM. Which dominated financial thought 00:11:43.460 --> 00:11:46.379 from the 60s onward. Right. And CAPM posited 00:11:46.379 --> 00:11:48.840 that the perfectly diversified investor should 00:11:48.840 --> 00:11:51.539 hold a pro -rata share of all available assets 00:11:51.539 --> 00:11:54.980 in the entire investment universe weighted by 00:11:54.980 --> 00:11:57.700 their market value. Okay, pro -rata share of 00:11:57.700 --> 00:12:00.559 all available assets. What does that translate 00:12:00.559 --> 00:12:02.899 to in practical terms for someone listening today? 00:12:03.159 --> 00:12:05.659 It translates directly to the modern index fund. 00:12:06.059 --> 00:12:08.360 When you invest in a fund that tracks the total 00:12:08.360 --> 00:12:11.320 global stock market, you are attempting to mimic 00:12:11.320 --> 00:12:13.940 this theoretical market portfolio. Trying to 00:12:13.940 --> 00:12:16.100 own a slice of everything. You're achieving maximum 00:12:16.100 --> 00:12:18.559 diversification within that particular asset 00:12:18.559 --> 00:12:21.679 class, global equity, because you own a tiny 00:12:21.679 --> 00:12:23.879 piece of everything weighted by how big that 00:12:23.879 --> 00:12:26.639 thing is in the global economy. That feels like 00:12:26.639 --> 00:12:29.639 the ultimate definition of dilution. If I buy 00:12:29.639 --> 00:12:32.220 the world, I can't possibly be destroyed by the 00:12:32.220 --> 00:12:34.620 failure of one specific company. That's the logic. 00:12:35.120 --> 00:12:38.080 And theoretically, diversification has no maximum 00:12:38.080 --> 00:12:40.120 so long as more unique assets are available. 00:12:40.360 --> 00:12:42.980 Every equally weighted, uncorrelated asset you 00:12:42.980 --> 00:12:45.419 add will mathematically contribute to reducing 00:12:45.419 --> 00:12:48.240 your portfolio's measured variance. But the real 00:12:48.240 --> 00:12:51.519 world is messy. As we saw in 2008 and during 00:12:51.519 --> 00:12:54.940 the pandemic, when a serious crisis hits, correlations 00:12:54.940 --> 00:12:57.500 tend to jump towards one. Everything starts moving 00:12:57.500 --> 00:13:00.220 down together. A very real problem. So how do 00:13:00.220 --> 00:13:02.419 the more advanced strategies account for assets 00:13:02.419 --> 00:13:05.419 that aren't uniformly uncorrelated? This is where 00:13:05.419 --> 00:13:08.039 practitioners move beyond just market cap weighting, 00:13:08.039 --> 00:13:10.379 which simply uses market size to determine your 00:13:10.379 --> 00:13:13.279 holding. We move into alternative risk based 00:13:13.279 --> 00:13:15.899 weighting approaches. And the first major strategy 00:13:15.899 --> 00:13:18.820 here is risk parity. OK, explain risk parity 00:13:18.820 --> 00:13:21.659 simply for us. Risk parity aims to weight assets 00:13:21.659 --> 00:13:24.019 in inverse proportion to their historical or 00:13:24.019 --> 00:13:26.539 their forecasted risk, which is usually measured 00:13:26.539 --> 00:13:28.779 by volatility. So if stocks are twice as volatile 00:13:28.779 --> 00:13:31.789 as bonds. A risk parity strategy would dictate 00:13:31.789 --> 00:13:34.710 holding twice as much value in bonds as in stocks. 00:13:34.929 --> 00:13:37.710 Wait. So instead of measuring my portfolio's 00:13:37.710 --> 00:13:40.110 contribution by dollar value, I'm measuring its 00:13:40.110 --> 00:13:43.409 contribution by risk. Exactly. The result is 00:13:43.409 --> 00:13:46.210 a portfolio where every asset class, be it commodities, 00:13:46.389 --> 00:13:49.409 stocks, bonds, real estate, contributes an equal 00:13:49.409 --> 00:13:51.669 amount of risk to the overall portfolio volatility. 00:13:52.210 --> 00:13:54.129 And why do it that way? What's the justification? 00:13:54.549 --> 00:13:57.690 It's highly pragmatic. Forecasting future risk 00:13:57.690 --> 00:14:00.730 or volatility is statistically much easier and 00:14:00.730 --> 00:14:03.230 more reliable than forecasting future market 00:14:03.230 --> 00:14:06.070 prices or economic growth. So you're structuring 00:14:06.070 --> 00:14:09.210 the portfolio around a known quantity risk contribution 00:14:09.210 --> 00:14:11.929 rather than an unknown quantity future return. 00:14:12.190 --> 00:14:14.269 That's a fascinating inversion of the typical 00:14:14.269 --> 00:14:16.690 approach, which is always chasing returns. This 00:14:16.690 --> 00:14:19.049 chases stability. And this leads to an even more 00:14:19.049 --> 00:14:22.289 nuanced approach, correlation parity. This is 00:14:22.289 --> 00:14:23.970 a further refinement. Instead of just aiming 00:14:23.970 --> 00:14:26.350 for equal volatility contribution, like in risk 00:14:26.350 --> 00:14:28.730 parity, correlation parity tries to structure 00:14:28.730 --> 00:14:31.529 the portfolio so that each individual asset have 00:14:31.529 --> 00:14:34.210 an equal correlation with the overall portfolio 00:14:34.210 --> 00:14:37.809 itself. So how does balancing correlation beat 00:14:37.809 --> 00:14:40.950 balancing volatility? Well, think of volatility 00:14:40.950 --> 00:14:43.649 as the speed of a car. And correlation is how 00:14:43.649 --> 00:14:45.649 many other cars on the road follow your exact 00:14:45.649 --> 00:14:48.789 speed changes. OK. Risk parity ensures every 00:14:48.789 --> 00:14:50.769 part of your portfolio is equally responsible 00:14:50.769 --> 00:14:53.690 for the overall speed bumps. Correlation parity 00:14:53.690 --> 00:14:56.269 tries to make sure every part is equally independent 00:14:56.269 --> 00:14:59.149 of the overall portfolio's movement. In mathematical 00:14:59.149 --> 00:15:02.169 terms, it's considered the true most diversified 00:15:02.169 --> 00:15:05.230 portfolio because it aims to reduce shared dependency 00:15:05.230 --> 00:15:08.149 across the board. So risk parity is just a special 00:15:08.149 --> 00:15:10.330 case of it. It's a special case of correlation 00:15:10.330 --> 00:15:13.549 parity that only applies if you assume all the 00:15:13.549 --> 00:15:16.990 correlations are uniform. These complex methods 00:15:16.990 --> 00:15:19.009 exist because they recognize that correlation 00:15:19.009 --> 00:15:21.690 fundamentally is the chief enemy of effective 00:15:21.690 --> 00:15:23.870 diversification. OK, let's bring it back to the 00:15:23.870 --> 00:15:26.830 CAPM model because it is so crucial for our listener 00:15:26.830 --> 00:15:29.710 to internalize the distinction the model introduced. 00:15:29.970 --> 00:15:32.450 We have to define the two faces of risk clearly. 00:15:32.629 --> 00:15:35.190 This separation is the true genius of modern 00:15:35.190 --> 00:15:37.799 portfolio theory. The first type of risk is the 00:15:37.799 --> 00:15:41.019 one we can get rid of, diversifiable risk. It's 00:15:41.019 --> 00:15:43.700 also called idiosyncratic risk, unsystematic 00:15:43.700 --> 00:15:46.840 risk, or security -specific risk. Give us a concrete 00:15:46.840 --> 00:15:49.460 example of diversifiable risk. Let's say you 00:15:49.460 --> 00:15:52.759 own shares in Exxon Marble. If a specific oil 00:15:52.759 --> 00:15:55.899 rig malfunctions or they lose a patent on a specific 00:15:55.899 --> 00:15:59.980 drilling technology, that incident only impacts 00:15:59.980 --> 00:16:02.419 ExxonMobil. It's specific to that company. Right. 00:16:02.519 --> 00:16:04.820 And because it's only one company in the S &P 00:16:04.820 --> 00:16:07.779 500, you can effectively dilute or eliminate 00:16:07.779 --> 00:16:10.200 the financial impact of that one event by holding 00:16:10.200 --> 00:16:13.440 the other 499 stocks. And the risk we absolutely 00:16:13.440 --> 00:16:15.919 cannot eliminate no matter how clever we get. 00:16:16.039 --> 00:16:18.730 That is the non -diversifiable risk. We also 00:16:18.730 --> 00:16:21.450 call this systematic risk or market risk, and 00:16:21.450 --> 00:16:24.049 it's often quantified by an asset's beta. So 00:16:24.049 --> 00:16:25.990 if you own an index fund tracking the entire 00:16:25.990 --> 00:16:28.610 S &P 500. You are still exposed to movements 00:16:28.610 --> 00:16:30.870 of that index as a whole. No matter how many 00:16:30.870 --> 00:16:32.710 individual stocks you buy within that market, 00:16:32.870 --> 00:16:35.210 you cannot eliminate the risk that the entire 00:16:35.210 --> 00:16:37.289 market dives due to a coordinated recession, 00:16:37.610 --> 00:16:40.190 a sudden spike in inflation or a global war. 00:16:40.639 --> 00:16:43.080 Because that risk applies to all companies, more 00:16:43.080 --> 00:16:45.580 or less. Directionally, yes. It applies to all 00:16:45.580 --> 00:16:48.399 of them. So the TAPM's argument is essentially, 00:16:48.559 --> 00:16:51.259 stop worrying about the risk you can control. 00:16:51.379 --> 00:16:54.360 Focus on the risk you can't. Absolutely. And 00:16:54.360 --> 00:16:57.500 they made an extremely powerful, logical leap 00:16:57.500 --> 00:17:00.399 about compensation. They argued that investors 00:17:00.399 --> 00:17:03.000 should only be financially compensated, meaning 00:17:03.000 --> 00:17:06.099 they should only expect a premium return for 00:17:06.099 --> 00:17:09.240 bearing non -diversifiable risk. And the logic 00:17:09.240 --> 00:17:13.200 there is? unassailable. Diversifiable risk is 00:17:13.200 --> 00:17:15.059 entirely within your control as an investor. 00:17:15.460 --> 00:17:17.880 If you choose to hold only one stock and you 00:17:17.880 --> 00:17:20.500 suffer a 50 percent loss due to a management 00:17:20.500 --> 00:17:23.160 scandal, the market views that as a self -inflicted 00:17:23.160 --> 00:17:25.660 wound. The market doesn't pay you extra just 00:17:25.660 --> 00:17:28.559 for failing to diversify. It does not. It creates 00:17:28.559 --> 00:17:30.500 an efficient market structure where unnecessary 00:17:30.500 --> 00:17:33.039 risk is not rewarded. Which puts the onus entirely 00:17:33.039 --> 00:17:35.740 on the investor. It does. But, you know, in this 00:17:35.740 --> 00:17:38.019 pursuit of maximum diversification, we have to 00:17:38.019 --> 00:17:41.079 introduce a necessary practical caveat. of over 00:17:41.079 --> 00:17:43.680 diversifying. Over diversifying. How can too 00:17:43.680 --> 00:17:46.400 much of a good thing possibly be bad? Well, while 00:17:46.400 --> 00:17:49.240 mathematically every uncorrelated asset helps, 00:17:49.480 --> 00:17:52.319 practically diversification has to be cost efficient. 00:17:52.619 --> 00:17:55.279 If you start adding assets where the per asset 00:17:55.279 --> 00:17:58.579 investment fees, the transaction costs or just 00:17:58.579 --> 00:18:01.339 the increased monitoring expenses begin to outweigh 00:18:01.339 --> 00:18:04.660 the marginal tiny gains you get from more diversification. 00:18:05.019 --> 00:18:07.759 Your net portfolio performance will suffer. Exactly. 00:18:07.819 --> 00:18:10.099 That was a much bigger problem, say, 30 years 00:18:10.099 --> 00:18:12.819 ago when mutual fund fees were high and trading 00:18:12.819 --> 00:18:15.619 was expensive. For sure. But even today. if you're 00:18:15.619 --> 00:18:18.640 actively trading small amounts of complex, illiquid 00:18:18.640 --> 00:18:21.859 assets, those fees or just the difficulty of 00:18:21.859 --> 00:18:25.099 managing a massive number of positions can quickly 00:18:25.099 --> 00:18:27.640 consume that marginal risk benefit. It's a pragmatic 00:18:27.640 --> 00:18:30.140 warning. It's a warning that the pure quantitative 00:18:30.140 --> 00:18:33.059 models sometimes overlook. The theoretical pursuit 00:18:33.059 --> 00:18:35.660 of maximum diversification must always be timbered 00:18:35.660 --> 00:18:37.920 by the practical costs of execution. All right, 00:18:37.940 --> 00:18:39.599 let's move beyond the theory and get into the 00:18:39.599 --> 00:18:41.700 engine room. We've claimed that diversification 00:18:41.700 --> 00:18:44.299 lowers variance even if assets aren't perfectly 00:18:44.490 --> 00:18:46.829 opposed. How does the math actually make this 00:18:46.829 --> 00:18:48.849 magic happen? We have to look at the mathematical 00:18:48.849 --> 00:18:51.109 structure of variance, which is our formal measure 00:18:51.109 --> 00:18:54.869 of risk, sigma squared. The beauty of diversification 00:18:54.869 --> 00:18:57.210 lies in the fact that when you combine assets, 00:18:57.569 --> 00:19:00.190 you're including both their individual risk contributions 00:19:00.190 --> 00:19:03.329 and, critically, their shared risk relationship, 00:19:03.630 --> 00:19:05.809 the covariance. Let's start with the base case 00:19:05.809 --> 00:19:09.609 again. Two perfectly uncorrelated assets, X and 00:19:09.609 --> 00:19:12.920 Y. Zero covariance. So we use our simple thought 00:19:12.920 --> 00:19:16.579 experiment. A total portfolio of, say, W dollars. 00:19:16.680 --> 00:19:20.380 You invest a fraction Q into asset X and the 00:19:20.380 --> 00:19:23.420 remaining fraction 1 minus Q into asset Y. Because 00:19:23.420 --> 00:19:26.299 we assumed zero correlation, the formula for 00:19:26.299 --> 00:19:29.599 the portfolio variance, sigma P22, it simplifies 00:19:29.599 --> 00:19:33.180 to 2Q dollars sigma X2 plus 1Q2 sigma E22. And 00:19:33.180 --> 00:19:35.180 I see the crucial element right there. The portfolio 00:19:35.180 --> 00:19:37.579 variance is the weighted sum of the individual 00:19:37.579 --> 00:19:40.769 variances. But the weights... The Q and the 1 00:19:40.769 --> 00:19:42.829 minus 4 squared. They're squared. And since Q 00:19:42.829 --> 00:19:45.470 and 1 minus Q are both less than 1, their squares 00:19:45.470 --> 00:19:47.730 are even smaller. Right. If Q is a half, 50 -50 00:19:47.730 --> 00:19:50.359 weighting, then Q squared is a quarter. Precisely. 00:19:50.380 --> 00:19:53.519 So instead of each asset contributing 50 % of 00:19:53.519 --> 00:19:56.720 its risk, it contributes only 25 % of its risk 00:19:56.720 --> 00:19:59.759 to the total variance pool. It's the squaring 00:19:59.759 --> 00:20:02.160 of the weights that creates this geometric reduction 00:20:02.160 --> 00:20:05.799 in risk. And furthermore, mathematically, you 00:20:05.799 --> 00:20:08.559 can always find a specific variance minimizing 00:20:08.559 --> 00:20:12.799 weighting, a Q, that is strictly between 0 and 00:20:12.799 --> 00:20:15.140 1. Which proves that the most stable portfolio 00:20:15.140 --> 00:20:17.940 requires holding both assets. It demonstrates 00:20:17.940 --> 00:20:20.619 that diversification... reduces risk below the 00:20:20.619 --> 00:20:23.339 level of even the least volatile single asset. 00:20:23.519 --> 00:20:26.460 That's the core mathematical proof. Even if the 00:20:26.460 --> 00:20:28.799 assets offer zero benefit from offsetting movements, 00:20:29.099 --> 00:20:32.150 just subdividing your capital across them immediately 00:20:32.150 --> 00:20:34.609 kills a portion of the total risk. And this favorable 00:20:34.609 --> 00:20:37.029 effect is, of course, enhanced exponentially 00:20:37.029 --> 00:20:39.609 if the assets are negatively correlated. One 00:20:39.609 --> 00:20:41.650 goes up when the other goes down, creating a 00:20:41.650 --> 00:20:43.769 natural hedge. And it's diminished but not eliminated 00:20:43.769 --> 00:20:45.849 if they're positively correlated. Diminished 00:20:45.849 --> 00:20:47.849 but still present, assuming the correlation is 00:20:47.849 --> 00:20:50.150 not perfectly positive, not equal to plus one. 00:20:50.369 --> 00:20:53.150 So let's scale this up dramatically. What happens 00:20:53.150 --> 00:20:56.089 when we look at N assets, all equally weighted, 00:20:56.230 --> 00:20:59.039 mutually uncorrelated? And they all have the 00:20:59.039 --> 00:21:02.079 same individual variance, sigma x222. The resulting 00:21:02.079 --> 00:21:04.920 portfolio variance simplifies to this elegantly 00:21:04.920 --> 00:21:08.859 clear expression, sigma p22 equals sigma x22 00:21:08.859 --> 00:21:12.079 divided by n. That is beautiful in its simplicity. 00:21:12.589 --> 00:21:15.309 It is. The portfolio variance is simply the individual 00:21:15.309 --> 00:21:17.869 asset variance divided by the number of assets 00:21:17.869 --> 00:21:21.569 you hold. This result is mathematically monotonically 00:21:21.569 --> 00:21:24.470 decreasing, as in increases. Meaning it only 00:21:24.470 --> 00:21:27.089 ever goes in one direction. Downwards. As you 00:21:27.089 --> 00:21:29.490 add more and more uncorrelated assets, the risk 00:21:29.490 --> 00:21:31.950 only ever goes down. If you start with 10 assets 00:21:31.950 --> 00:21:34.009 and you double your holdings to 20, you cut the 00:21:34.009 --> 00:21:36.430 variance in half. This is the clearest possible 00:21:36.430 --> 00:21:39.109 mathematical statement on the power of diversification. 00:21:39.549 --> 00:21:41.849 But we need to spend a moment on a critical technical... 00:21:41.869 --> 00:21:44.329 distinction that financial theory demands. The 00:21:44.329 --> 00:21:46.910 difference between true diversification and simply 00:21:46.910 --> 00:21:48.589 adding risk. You brought up the example of an 00:21:48.589 --> 00:21:51.589 insurance company. Yes. When we talk about sigma 00:21:51.589 --> 00:21:54.250 squared divided by n, we are defining diversification 00:21:54.250 --> 00:21:57.210 as subdividing a fixed total amount of capital. 00:21:57.430 --> 00:22:00.630 We have one pie and we slice it into n pieces. 00:22:00.890 --> 00:22:03.309 Right. Now consider the opposite scenario as 00:22:03.309 --> 00:22:06.069 explored by thinkers like Paul Samuelson. What 00:22:06.069 --> 00:22:08.329 if an entity is simply aggregating independent 00:22:08.329 --> 00:22:11.029 risks? Like an insurance company adding policy 00:22:11.029 --> 00:22:14.289 X1, then policy X2, then X3, and so on? Exactly. 00:22:14.490 --> 00:22:17.849 Each policy is an uncorrelated volatile risk. 00:22:18.069 --> 00:22:20.869 If the insurance company simply adds N uncorrelated 00:22:20.869 --> 00:22:23.569 policies without changing its capital base, the 00:22:23.569 --> 00:22:26.269 total variance of their exposure is N times sigma 00:22:26.269 --> 00:22:28.569 squared. The variance is increasing. It's increasing 00:22:28.569 --> 00:22:31.930 linearly with N. They are adding risk, not diluting 00:22:31.930 --> 00:22:34.670 it. That's a key conceptual barrier. The insurance 00:22:34.670 --> 00:22:36.630 company... becomes riskier, the more policies 00:22:36.630 --> 00:22:39.150 it writes, all else being equal. The company 00:22:39.150 --> 00:22:42.150 aggregates the risk. The diversification in this 00:22:42.150 --> 00:22:44.309 case doesn't happen at the corporate level by 00:22:44.309 --> 00:22:46.809 writing more policies. It happens when that risk 00:22:46.809 --> 00:22:49.210 is ultimately dispersed among the company's investors, 00:22:49.410 --> 00:22:51.730 the shareholders who spread their investment 00:22:51.730 --> 00:22:53.910 across many companies and other asset types. 00:22:54.089 --> 00:22:57.170 The company aggregates risk. The investor dilutes 00:22:57.170 --> 00:22:59.109 it. Okay, that's the theory. That's the mathematical 00:22:59.109 --> 00:23:01.750 scaffolding. Let's make this palpable for the 00:23:01.750 --> 00:23:03.710 listener by looking at the classic empirical 00:23:03.710 --> 00:23:06.829 proof, the study that quantified exactly how 00:23:06.829 --> 00:23:09.430 fast this risk falls off. We're talking about 00:23:09.430 --> 00:23:12.619 the 1977 work by Elton and Gruber. This study 00:23:12.619 --> 00:23:15.079 is the bedrock of the how many stocks are enough 00:23:15.079 --> 00:23:18.079 debate. They looked at a huge population of available 00:23:18.079 --> 00:23:21.220 securities and calculated the average risk, the 00:23:21.220 --> 00:23:23.960 standard deviation of annual returns over all 00:23:23.960 --> 00:23:27.160 possible randomly chosen portfolios of size N 00:23:27.160 --> 00:23:30.140 with an equal amount held in each asset. The 00:23:30.140 --> 00:23:32.500 power here is that it wasn't theoretical. It 00:23:32.500 --> 00:23:35.039 was based on historical market data. So let's 00:23:35.039 --> 00:23:36.500 look at the steepness of that risk reduction 00:23:36.500 --> 00:23:39.720 curve. For a portfolio with just one stock, they 00:23:39.720 --> 00:23:43.420 found the average standard deviation was 49 .24%. 00:23:43.420 --> 00:23:46.839 That is maximum idiosyncratic risk. We'll use 00:23:46.839 --> 00:23:50.019 that as our baseline, a ratio of one. Now watch 00:23:50.019 --> 00:23:52.680 the immediate return on diversification. How 00:23:52.680 --> 00:23:54.740 much risk is killed by just adding three more 00:23:54.740 --> 00:23:57.240 stocks? With four stocks, the standard deviation 00:23:57.240 --> 00:24:01.700 plummeted to 29 .69%. That's a ratio of 0 .60. 00:24:01.900 --> 00:24:04.839 So you eliminated 40 % of the total measurable 00:24:04.839 --> 00:24:08.009 risk just by going from one to four. 40%. That 00:24:08.009 --> 00:24:10.210 is the massive initial gain we were talking about. 00:24:10.369 --> 00:24:12.329 And if we move up to 20 stocks, which is that 00:24:12.329 --> 00:24:14.789 comfortable threshold for eliminating idiosyncratic 00:24:14.789 --> 00:24:17.470 risk. At 20 stocks, the standard deviation was 00:24:17.470 --> 00:24:22.450 21 .68%, a risk ratio of 0 .44. So 20 random 00:24:22.450 --> 00:24:24.670 stocks reduced the risk to less than half of 00:24:24.670 --> 00:24:26.829 what a single stock entailed. And moving from 00:24:26.829 --> 00:24:30.160 20 to the historical benchmark of 30. At 30 stocks, 00:24:30.440 --> 00:24:33.920 the standard deviation was 20 .87%, a ratio of 00:24:33.920 --> 00:24:36.779 0 .42. The incremental benefit of those extra 00:24:36.779 --> 00:24:39.839 10 stocks was minimal. Just 2 % more risk reduction. 00:24:40.079 --> 00:24:42.380 Tiny. And here's the truly astonishing part. 00:24:42.519 --> 00:24:45.079 Let's compare that 30 -stock figure to a maximally 00:24:45.079 --> 00:24:48.259 diversified holding of 1 ,000 stocks. At 1 ,000 00:24:48.259 --> 00:24:50.680 stocks, the risk reduction essentially flatlines. 00:24:50.839 --> 00:24:53.400 It drops only slightly further to a standard 00:24:53.400 --> 00:24:58.200 deviation of 19 .21%, a ratio of 0 .39. The ultimate 00:24:58.200 --> 00:25:00.119 takeaway for the listener is just undeniable 00:25:00.119 --> 00:25:03.079 here. Four stocks capture nearly half the benefit. 00:25:03.480 --> 00:25:05.900 And the risk level you get with 30 stocks is 00:25:05.900 --> 00:25:07.859 extremely close to the risk level of holding 00:25:07.859 --> 00:25:11.319 1 ,000. The curve flattens out so rapidly. Beyond 00:25:11.319 --> 00:25:14.740 that point, that residual risk, that 39%, that 00:25:14.740 --> 00:25:17.920 is the non -diversifiable systematic risk. It 00:25:17.920 --> 00:25:19.920 confirms that your effort should be focused on 00:25:19.920 --> 00:25:23.119 building that core diversified portfolio of 15 00:25:23.119 --> 00:25:26.440 to 30 well -chosen non -correlated assets rather 00:25:26.440 --> 00:25:28.859 than inefficiently trying to replicate the market 00:25:28.859 --> 00:25:31.500 through thousands of individual high -cost trades. 00:25:32.809 --> 00:25:34.450 the quantitative rules for the individual industry, 00:25:34.650 --> 00:25:36.470 but these same concepts of risk and correlation, 00:25:36.630 --> 00:25:39.390 they play at the corporate level too. Large companies 00:25:39.390 --> 00:25:41.130 also have to manage risk, not just of single 00:25:41.130 --> 00:25:43.430 assets, but of entire product lines or geographical 00:25:43.430 --> 00:25:46.349 regions. Let's explore corporate diversification 00:25:46.349 --> 00:25:49.279 strategies. When corporations diversify, they're 00:25:49.279 --> 00:25:51.799 generally doing it for stability, either internally 00:25:51.799 --> 00:25:54.359 within their supply chain or externally against 00:25:54.359 --> 00:25:58.119 market volatility. The three common forms really 00:25:58.119 --> 00:26:00.700 just define their relationship to the core business. 00:26:01.039 --> 00:26:03.160 Okay, start with horizontal diversification. 00:26:03.640 --> 00:26:05.640 This is the most common and often the lowest 00:26:05.640 --> 00:26:08.140 risk approach. It involves expanding a product 00:26:08.140 --> 00:26:10.759 line that uses existing resources or acquiring 00:26:10.759 --> 00:26:13.680 companies that operate adjacent to the core business. 00:26:14.000 --> 00:26:15.960 So a soft drink company starts selling energy 00:26:15.960 --> 00:26:18.059 bars. Perfect example. They already have the 00:26:18.059 --> 00:26:19.940 distribution networks. They have the marketing 00:26:19.940 --> 00:26:22.220 muscle. They're just expanding the product offering 00:26:22.220 --> 00:26:25.710 horizontally. And vertical diversification. That 00:26:25.710 --> 00:26:28.130 sounds like capturing the supply chain. Precisely. 00:26:28.210 --> 00:26:31.109 Vertical integration is all about consolidating 00:26:31.109 --> 00:26:33.430 steps in the value chain amalgamating distribution 00:26:33.430 --> 00:26:36.869 channels, or integrating the supply chain. A 00:26:36.869 --> 00:26:39.589 large automaker might buy a lithium mining operation. 00:26:39.990 --> 00:26:41.809 Or a software firm that writes their in -car 00:26:41.809 --> 00:26:45.890 code. Exactly. The goal is efficiency, cost control, 00:26:46.069 --> 00:26:48.690 and securing critical inputs, which mitigates 00:26:48.690 --> 00:26:51.170 the risk of supply disruptions. And then the 00:26:51.170 --> 00:26:53.690 most aggressive form. one that has fallen in 00:26:53.690 --> 00:26:55.849 and out of favor dramatically over the decades, 00:26:56.069 --> 00:26:59.650 non -incremental diversification. This is the 00:26:59.650 --> 00:27:01.930 strategy of the pure conglomerate. It involves 00:27:01.930 --> 00:27:04.589 acquiring or launching business lines that have 00:27:04.589 --> 00:27:06.829 little to no connection to the company's core 00:27:06.829 --> 00:27:10.130 activities. A classic example would be a company 00:27:10.130 --> 00:27:12.549 that simultaneously owns aerospace manufacturing, 00:27:12.950 --> 00:27:15.829 a commercial real estate portfolio, and a media 00:27:15.829 --> 00:27:18.269 publishing house. The famous example being Berkshire 00:27:18.269 --> 00:27:20.730 Hathaway, although they'd argue their strategy 00:27:20.730 --> 00:27:23.349 is more about capital allocation than traditional 00:27:23.349 --> 00:27:26.569 operational synergy. Why adopt such a disconnected 00:27:26.569 --> 00:27:29.589 approach? The theory is purely diversification 00:27:29.589 --> 00:27:32.750 driven. It's to stabilize the overall company's 00:27:32.750 --> 00:27:35.509 earnings by mitigating external exogenous risk 00:27:35.509 --> 00:27:37.950 factors. The hope is that when the aerospace 00:27:37.950 --> 00:27:40.390 cycle is depressed, the commercial real estate 00:27:40.390 --> 00:27:42.869 market might be booming or vice versa. The different 00:27:42.869 --> 00:27:45.309 lines smooth out the total income flow. They 00:27:45.309 --> 00:27:47.750 provide stability to shareholders and often lower 00:27:47.750 --> 00:27:50.589 the overall cost of capital for the entity. Although 00:27:50.589 --> 00:27:53.289 the complexity of managing such disparate businesses 00:27:53.289 --> 00:27:55.769 is often what ultimately leads to conglomerates 00:27:55.769 --> 00:27:58.180 being broken up. It's corporate diversification 00:27:58.180 --> 00:28:01.359 applied on a grand scale. You're trying to kill 00:28:01.359 --> 00:28:04.640 the systematic risk of specific sectors by owning 00:28:04.640 --> 00:28:07.779 a basket of uncorrelated sectors. Exactly. And 00:28:07.779 --> 00:28:10.960 its success often relies entirely on management's 00:28:10.960 --> 00:28:13.299 ability to allocate capital effectively across 00:28:13.299 --> 00:28:16.380 those disparate businesses, not on operational 00:28:16.380 --> 00:28:19.009 synergy. Now, let's pivot to a point that you 00:28:19.009 --> 00:28:21.890 need to be very clear on because it impacts retirement 00:28:21.890 --> 00:28:24.529 planning for virtually every single person listening. 00:28:24.730 --> 00:28:27.910 The fallacy of time diversification. This is 00:28:27.910 --> 00:28:30.670 the belief that time itself is a risk reducer. 00:28:31.079 --> 00:28:34.559 This belief is so deeply rooted in investor psychology, 00:28:34.960 --> 00:28:38.099 it's taught widely in personal finance. The argument 00:28:38.099 --> 00:28:40.440 goes that because younger investors have decades 00:28:40.440 --> 00:28:43.140 until retirement, they should aggressively favor 00:28:43.140 --> 00:28:45.599 risky assets like stocks. Because they have time 00:28:45.599 --> 00:28:48.240 to recover from any downturns. Right. The underlying 00:28:48.240 --> 00:28:51.160 and very misleading premise is that over long 00:28:51.160 --> 00:28:53.400 periods, the good returns will mathematically 00:28:53.400 --> 00:28:56.599 cancel out any possible bad returns, making the 00:28:56.599 --> 00:28:58.839 long run safer. It's that comforting voice that 00:28:58.839 --> 00:29:01.119 says, Don't worry about the market crash today. 00:29:01.279 --> 00:29:03.700 You have 30 years left. And while that voice 00:29:03.700 --> 00:29:06.279 is attempting to calm panic, it fundamentally 00:29:06.279 --> 00:29:09.740 misrepresents how risk operates over time. The 00:29:09.740 --> 00:29:12.660 flaw was definitively identified by Nobel laureate 00:29:12.660 --> 00:29:15.119 Paul Samuelson, among others like John Norstad 00:29:15.119 --> 00:29:18.230 and Zvi Body. So what's the flaw? While it is 00:29:18.230 --> 00:29:20.490 true that the standard deviation of annualized 00:29:20.490 --> 00:29:23.349 returns decreases as the investment horizon increases, 00:29:23.549 --> 00:29:25.869 meaning your average yearly return becomes more 00:29:25.869 --> 00:29:28.930 stable, that is not the measure the retiree actually 00:29:28.930 --> 00:29:31.309 cares about. The retiree cares about the final 00:29:31.309 --> 00:29:34.319 dollar amount. the total return precisely the 00:29:34.319 --> 00:29:36.660 uncertainty that matters is the final outcome 00:29:36.660 --> 00:29:39.960 and due to the inexorable power of compounding 00:29:39.960 --> 00:29:42.619 the standard deviation of the total return the 00:29:42.619 --> 00:29:44.599 uncertainty surrounding your final accumulated 00:29:44.599 --> 00:29:47.119 wealth actually increases dramatically with the 00:29:47.119 --> 00:29:49.259 length of the time horizon let's visualize that 00:29:49.259 --> 00:29:52.180 if i invest for one year the range of possible 00:29:52.180 --> 00:29:54.539 outcomes from my final capital is relatively 00:29:54.539 --> 00:29:58.460 small Maybe I earn 10 % or I lose 10%. Correct. 00:29:58.859 --> 00:30:02.480 But if you invest for 40 years, the range of 00:30:02.480 --> 00:30:05.240 possible total returns due to the cumulative 00:30:05.240 --> 00:30:08.380 effect of those annual gains and losses compounded 00:30:08.380 --> 00:30:11.819 over four decades is exponentially larger. The 00:30:11.819 --> 00:30:14.660 absolute risk to your final nest egg is greater, 00:30:14.740 --> 00:30:18.259 not less, the longer the timeline. So uncertainty, 00:30:18.440 --> 00:30:21.019 measured as the standard deviation of my total 00:30:21.019 --> 00:30:24.529 dollar return, increases with time. It increases 00:30:24.529 --> 00:30:26.869 with time. So if I start with $10 ,000 and I 00:30:26.869 --> 00:30:29.549 have a 40 year horizon, my best case scenario 00:30:29.549 --> 00:30:32.410 might be $5 million, but my worst case scenario 00:30:32.410 --> 00:30:34.950 might still be $50 ,000 depending on the sequence 00:30:34.950 --> 00:30:36.829 of returns. And the spread between those two 00:30:36.829 --> 00:30:38.730 numbers is the standard deviation of your total 00:30:38.730 --> 00:30:41.529 return. And that spread is massive. It's huge. 00:30:41.730 --> 00:30:44.190 The fallacy grants investors a false sense of 00:30:44.190 --> 00:30:46.750 security. It encourages them to concentrate their 00:30:46.750 --> 00:30:49.470 portfolio and high risk assets based on the mistaken 00:30:49.470 --> 00:30:52.130 belief that time acts as a natural systematic 00:30:52.130 --> 00:30:55.559 hedge. It does not. It magnifies potential outcomes. 00:30:55.740 --> 00:30:57.319 So long -term investors need diversification 00:30:57.319 --> 00:30:59.880 across asset classes, stocks, bonds, real estate, 00:30:59.920 --> 00:31:02.220 just as much as short -term investors. Absolutely. 00:31:02.539 --> 00:31:05.519 Simply to manage the volatility of that massive 00:31:05.519 --> 00:31:08.440 final wealth number. That reframes the entire 00:31:08.440 --> 00:31:10.779 purpose of diversification for a young person. 00:31:11.130 --> 00:31:13.829 It's not about surviving the next quarter. It's 00:31:13.829 --> 00:31:16.670 about controlling the enormous uncertainty inherent 00:31:16.670 --> 00:31:19.589 in a multi -decade compounding experiment. It's 00:31:19.589 --> 00:31:21.849 really remarkable that an economic concept so 00:31:21.849 --> 00:31:24.589 mathematically complex in its modern form has 00:31:24.589 --> 00:31:28.769 such simple ancient roots. Diversification in 00:31:28.769 --> 00:31:31.150 its core form is just applied common sense about 00:31:31.150 --> 00:31:33.230 survival. So let's trace this back through time. 00:31:33.450 --> 00:31:35.930 Where does the concept first appear in written 00:31:35.930 --> 00:31:38.710 history as a piece of financial advice? We can 00:31:38.710 --> 00:31:40.809 trace the concept back to the Bible, specifically 00:31:40.809 --> 00:31:43.250 the book of Ecclesiastes, which is often dated 00:31:43.250 --> 00:31:46.450 around 935 BC. The text offers this remarkably 00:31:46.450 --> 00:31:49.569 relevant financial advice. But divide your investments 00:31:49.569 --> 00:31:52.230 among many places, for you do not know what risks 00:31:52.230 --> 00:31:55.009 might lie ahead. That is nearly 3 ,000 years 00:31:55.009 --> 00:31:57.910 of consensus. The underlying concern is identical 00:31:57.910 --> 00:32:00.710 to modern risk management, preparing for unknown 00:32:00.710 --> 00:32:04.200 future risks. Another ancient source provides 00:32:04.200 --> 00:32:06.880 a highly specific prescriptive asset allocation 00:32:06.880 --> 00:32:10.740 that is still studied today, the Talmud. It details 00:32:10.740 --> 00:32:13.140 a specific financial strategy for wealth management. 00:32:13.319 --> 00:32:15.740 What was the Talmudic formula? The Talmud advised 00:32:15.740 --> 00:32:19.000 splitting one's assets into thirds. One third 00:32:19.000 --> 00:32:22.119 was to be held in business, so active trading 00:32:22.119 --> 00:32:24.220 and commerce. That was your high -risk, high 00:32:24.220 --> 00:32:26.880 -growth engine. The second third was to be kept 00:32:26.880 --> 00:32:29.569 liquid. perhaps in the form of gold coins or 00:32:29.569 --> 00:32:32.589 readily available cash. This was the stable zero 00:32:32.589 --> 00:32:35.349 risk component. And the final third was to be 00:32:35.349 --> 00:32:37.789 invested in land or real estate, representing 00:32:37.789 --> 00:32:40.849 the long term inflation protected tangible asset. 00:32:41.309 --> 00:32:43.950 That is sophisticated risk management. You have 00:32:43.950 --> 00:32:46.490 a growth engine, a safety net for liquidity and 00:32:46.490 --> 00:32:48.869 a stable store of value protected against economic 00:32:48.869 --> 00:32:51.049 turmoil. And this approach is now recognized 00:32:51.049 --> 00:32:54.150 in modern research as naive diversification or 00:32:54.150 --> 00:32:56.930 one end diversification because it instructs 00:32:56.930 --> 00:32:59.150 the investor to split wealth equally among the 00:32:59.150 --> 00:33:01.349 available major asset classes. And it actually 00:33:01.349 --> 00:33:04.180 works. Research since the year 2000 has shown 00:33:04.180 --> 00:33:07.099 that, surprisingly, this simple 1N allocation 00:33:07.099 --> 00:33:10.319 strategy often performs remarkably well compared 00:33:10.319 --> 00:33:13.880 to much more complex, optimization -driven models, 00:33:14.079 --> 00:33:16.559 especially during periods of high market volatility. 00:33:17.019 --> 00:33:19.819 The ancient wisdom holds significant practical 00:33:19.819 --> 00:33:23.039 value. And we even see this concept woven into 00:33:23.039 --> 00:33:25.660 the fabric of Renaissance literature, reflecting 00:33:25.660 --> 00:33:28.200 how crucial it was to global trade back then. 00:33:28.619 --> 00:33:30.380 You just have to think of Shakespeare's Merchant 00:33:30.380 --> 00:33:33.720 of Venice, written around 1599. The character 00:33:33.720 --> 00:33:36.619 Antonio, whose entire fortune is tied up in global 00:33:36.619 --> 00:33:39.619 shipping, states that his ventures are not in 00:33:39.619 --> 00:33:42.660 one bottom trusted, nor to one place. He's talking 00:33:42.660 --> 00:33:44.480 about ships and ports. Bottoms and ports, exactly. 00:33:44.940 --> 00:33:47.240 He fully understands that he has to mitigate 00:33:47.240 --> 00:33:50.319 the risk of piracy, storms, or a single port 00:33:50.319 --> 00:33:52.640 closure by spreading his ventures across multiple 00:33:52.640 --> 00:33:54.880 vessels and geographic routes at the same time. 00:33:55.039 --> 00:33:57.140 So the principle was intuitive for millennia. 00:33:57.390 --> 00:33:59.970 driven by the realities of physical risk. But 00:33:59.970 --> 00:34:02.349 the move towards scientific quantitative measurement, 00:34:02.549 --> 00:34:05.390 the transition to modern portfolio theory, that's 00:34:05.390 --> 00:34:08.000 where Harry Markovits steps in. Markovits' seminal 00:34:08.000 --> 00:34:10.699 work in the 1950s provided the necessary mathematical 00:34:10.699 --> 00:34:13.780 language. He moved the conversation from don't 00:34:13.780 --> 00:34:16.539 put all your eggs in one basket to we must calculate 00:34:16.539 --> 00:34:18.900 the covariance between the returns of the eggs 00:34:18.900 --> 00:34:21.940 to achieve optimal risk reduction. MPT. Modern 00:34:21.940 --> 00:34:24.820 portfolio theory. It formally defined how to 00:34:24.820 --> 00:34:26.960 get the highest possible return for a given level 00:34:26.960 --> 00:34:29.840 of risk or conversely the lowest possible risk 00:34:29.840 --> 00:34:32.360 for a target return. But even before Markovits 00:34:32.360 --> 00:34:35.739 formalized MPT, our sources highlight a major 00:34:35.739 --> 00:34:38.130 financial figure who was practicing sophisticated, 00:34:38.650 --> 00:34:41.449 intuitively driven diversification decades earlier. 00:34:41.929 --> 00:34:44.769 John Maynard Keynes. Keynes' tenure managing 00:34:44.769 --> 00:34:47.110 the King's College Cambridge endowment from the 00:34:47.110 --> 00:34:51.030 1920s to 1946 is legendary. He was, by modern 00:34:51.030 --> 00:34:53.170 standards, sometimes highly concentrated in his 00:34:53.170 --> 00:34:55.250 stock picks. But his underlying philosophy was 00:34:55.250 --> 00:34:57.510 profoundly diversified in the right ways. So 00:34:57.510 --> 00:34:59.670 what defined his risk management philosophy in 00:34:59.670 --> 00:35:02.150 a time without modern derivatives or index funds? 00:35:02.530 --> 00:35:04.449 He was focused on the structural relationship 00:35:04.449 --> 00:35:07.349 between assets. He recognized the crucial importance 00:35:07.349 --> 00:35:09.769 of holding assets with opposed risks. Meaning 00:35:09.769 --> 00:35:11.550 things that were likely to move and opposite 00:35:11.550 --> 00:35:14.590 directions exactly assets that would move in 00:35:14.590 --> 00:35:17.030 opposite directions during general market fluctuations 00:35:17.030 --> 00:35:20.150 he was intuitively grasping the power of low 00:35:20.150 --> 00:35:23.329 or negative correlation the key mathematical 00:35:23.329 --> 00:35:26.690 insight markovits later formalized he understood 00:35:26.690 --> 00:35:29.150 that while an individual asset might be volatile 00:35:29.150 --> 00:35:31.750 if you paired it correctly the portfolio could 00:35:31.750 --> 00:35:34.050 be stable and he was a pioneer in a geographic 00:35:34.050 --> 00:35:36.469 sense too he was avoiding that infamous home 00:35:36.469 --> 00:35:40.070 bias oh absolutely Keynes was a true maverick 00:35:40.070 --> 00:35:42.829 in this regard. He aggressively invested internationally, 00:35:43.110 --> 00:35:45.030 holding significant portions of the portfolio, 00:35:45.389 --> 00:35:48.510 sometimes as high as 75 percent in non -U .K. 00:35:48.590 --> 00:35:51.510 stocks, primarily U .S. equities. Which was revolutionary 00:35:51.510 --> 00:35:53.949 at the time. It was. In the 20s and 30s, nearly 00:35:53.949 --> 00:35:56.469 all university endowments in both the U .S. and 00:35:56.469 --> 00:35:59.730 the U .K. were heavily, often entirely, invested 00:35:59.730 --> 00:36:03.070 in domestic bonds and local assets. Keynes recognized 00:36:03.070 --> 00:36:05.130 that the risks facing the British economy were 00:36:05.130 --> 00:36:08.579 idiosyncratic on a global scale. He used international 00:36:08.579 --> 00:36:11.199 diversification as a core risk mitigation tool 00:36:11.199 --> 00:36:13.880 decades before it became institutional standard 00:36:13.880 --> 00:36:18.159 practice. That global perspective, based on observing 00:36:18.159 --> 00:36:20.480 risk factors rather than just relying on local 00:36:20.480 --> 00:36:23.639 comfort, really cements his position as a crucial 00:36:23.639 --> 00:36:26.539 historical figure in risk management. Hashtag 00:36:26.539 --> 00:36:29.599 tag outro. We've taken a very deep dive into 00:36:29.599 --> 00:36:32.519 this foundational principle. Let's try and synthesize 00:36:32.519 --> 00:36:34.260 the core knowledge points that you can take away 00:36:34.260 --> 00:36:37.000 right now. First and foremost, diversification 00:36:37.000 --> 00:36:39.760 is a risk management tool. That's it. It's designed 00:36:39.760 --> 00:36:42.219 to stabilize your returns, dramatically narrow 00:36:42.219 --> 00:36:45.280 the range of possible outcomes. It is not a mechanism 00:36:45.280 --> 00:36:48.000 for boosting your expected average returns. Second, 00:36:48.159 --> 00:36:50.099 the empirical evidence that Elton and Gruber 00:36:50.099 --> 00:36:53.019 study, it gives you a clear, actionable threshold. 00:36:53.320 --> 00:36:55.860 The most impactful risk reduction happens very 00:36:55.860 --> 00:36:59.309 quickly. Holding 15 to 20 well -chosen non -correlated 00:36:59.309 --> 00:37:01.929 assets captures the vast majority of the benefit 00:37:01.929 --> 00:37:04.389 available. It eliminates company -specific risk 00:37:04.389 --> 00:37:06.969 with high efficiency. And finally, ditch the 00:37:06.969 --> 00:37:09.190 comforting but misleading advice about time. 00:37:09.389 --> 00:37:11.789 Time does not reduce the uncertainty of your 00:37:11.789 --> 00:37:14.250 total final wealth. Due to compounding effects, 00:37:14.730 --> 00:37:17.389 a longer time horizon actually increases the 00:37:17.389 --> 00:37:20.110 standard deviation of your total return. This 00:37:20.110 --> 00:37:22.750 demands that long -term investors still rely 00:37:22.750 --> 00:37:25.130 heavily on diversification across asset classes 00:37:25.130 --> 00:37:28.449 to manage that enormous potential spread between 00:37:28.449 --> 00:37:30.889 their best and worst case retirement scenarios. 00:37:31.469 --> 00:37:33.929 And to leave you with one final provocative thought 00:37:33.929 --> 00:37:36.369 that circles back to where we started with systematic 00:37:36.369 --> 00:37:39.269 risk, let's just recall the asymptotic limit 00:37:39.269 --> 00:37:42.050 of diversification. We established that as the 00:37:42.050 --> 00:37:45.769 number of assets n approaches infinity in a maximized, 00:37:45.769 --> 00:37:48.289 equally weighted portfolio, the portfolio variance 00:37:48.289 --> 00:37:50.429 does not go to zero. It bottoms out. It hits 00:37:50.429 --> 00:37:53.780 a floor. Exactly. The variance tends asymptotically 00:37:53.780 --> 00:37:56.199 toward the average of the covariances between 00:37:56.199 --> 00:37:58.559 all the securities, what we call dom or sigma 00:37:58.559 --> 00:38:01.280 d. This shared risk, this average relatedness 00:38:01.280 --> 00:38:04.119 of all the assets, that is the ultimate non -diversifiable 00:38:04.119 --> 00:38:06.539 risk. The systematic risk that is woven into 00:38:06.539 --> 00:38:08.619 the fabric of the entire market. It's the risk 00:38:08.619 --> 00:38:10.900 that's left over. So if mathematically, even 00:38:10.900 --> 00:38:13.420 infinite diversification cannot eliminate the 00:38:13.420 --> 00:38:16.340 risk shared by every asset in the economy, what 00:38:16.340 --> 00:38:17.980 does that tell you about truly safe investing? 00:38:18.360 --> 00:38:20.739 How can an investor truly escape systematic risk 00:38:20.739 --> 00:38:22.820 if the entire global financial system is wired 00:38:22.820 --> 00:38:24.639 together by shared economic and geopolitical 00:38:24.639 --> 00:38:27.119 factors? That is the ultimate challenge for any 00:38:27.119 --> 00:38:28.699 serious investor to contemplate.