1 00:00:01,040 --> 00:00:05,484 You're about to join Niels Kaastrup-Larsen on a raw and honest 2 00:00:05,484 --> 00:00:09,524 journey into the world of systematic investing and learn about 3 00:00:09,524 --> 00:00:13,160 the most dependable and consistent yet often overlooked investment 4 00:00:13,160 --> 00:00:15,988 strategy. Welcome to the Systematic Investor series. 5 00:00:23,440 --> 00:00:26,490 Welcome back to the latest edition of Top Traders Unplugged 6 00:00:26,490 --> 00:00:30,455 where each week we take the pulse of the markets from the perspective 7 00:00:30,455 --> 00:00:33,810 of a rules-based investor. It's Alan Dunne here sitting in for 8 00:00:33,810 --> 00:00:36,860 Niels, joined by Mark this week. Mark, how are you? 9 00:00:37,340 --> 00:00:42,086 Not too bad. I was just in Chicago the last couple of days in 10 00:00:42,086 --> 00:00:45,476 60 degrees Fahrenheit which for Chicago, in February, is quite 11 00:00:45,476 --> 00:00:45,815 unusual. 12 00:00:46,620 --> 00:00:49,680 Very nice. Well, I'm in Dublin. It's been raining nonstop 13 00:00:49,680 --> 00:00:53,658 for, I don't know, at least two months it feels like. So, I'm 14 00:00:53,658 --> 00:00:57,330 off to Miami next week. Niels is already out there. Obviously we've 15 00:00:57,330 --> 00:01:01,002 got a big iConnections hedge fund conference, so, a lot of people 16 00:01:01,002 --> 00:01:04,368 will be out there. So, if you're there, say hello. Looking 17 00:01:04,368 --> 00:01:06,816 forward to that. And obviously, yeah, hopefully we'll 18 00:01:06,816 --> 00:01:10,488 see… No doubt, we'll see some sunshine in Miami, you would expect 19 00:01:10,488 --> 00:01:12,018 at this time of year. 20 00:01:12,040 --> 00:01:16,499 I'm going to say that there's so many hedge fund managers in Miami, 21 00:01:16,499 --> 00:01:20,272 Florida this time of year, it's almost as though, who's minding 22 00:01:20,272 --> 00:01:22,673 to shop if everybody's down in Florida? 23 00:01:23,320 --> 00:01:30,855 That's right. Well, yeah, hopefully it'll be a good week. I'm 24 00:01:30,855 --> 00:01:38,390 sure it will. Lots of meetings and it'll be very busy. Wealways 25 00:01:24,826 --> 00:01:33,862 start off with what's on your radar. Anything out of the ordinary 26 00:01:33,862 --> 00:01:38,380 catching your attention at the moment? 27 00:01:38,790 --> 00:01:43,132 Well, everything seems to be out of the ordinary in the last month, 28 00:01:43,132 --> 00:01:46,806 since the beginning of the year, between all of the bubbles 29 00:01:46,806 --> 00:01:50,480 we've had the craziness in geopolitics. So where do we begin? 30 00:01:50,480 --> 00:01:51,482 Hard to say. 31 00:01:52,950 --> 00:02:03,800 Fair enough. Well, we'll get to all of that in a minute. I think 32 00:02:03,800 --> 00:02:13,100 it makes sense to cover on performance, before we get into the 33 00:02:13,100 --> 00:02:21,625 main topics. So, on the month, trend following and managed futures 34 00:02:21,625 --> 00:02:29,375 continuing its recent good streak of performance. Month to date, 35 00:02:29,375 --> 00:02:38,675 the SocGen CTA index up 1.1%. And the SocGen Trend index slightly 36 00:02:38,675 --> 00:02:48,750 better at 1.15%, and then year to date pretty much neck and neck, 37 00:02:48,750 --> 00:02:58,825 both the SocGen CTA index and the SocGen Trend index up 5.9% on 38 00:02:58,825 --> 00:03:07,350 the year. So, it's obviously been a good period this month. Obviously,we've 39 00:01:54,500 --> 00:02:05,350 seen a bit of a dip down in equities but continue to see some 40 00:02:05,350 --> 00:02:13,100 decent moves on the commodities side, I would say. Obviously 41 00:02:13,100 --> 00:02:23,950 gold is higher on the month, and that's been a big trend in the 42 00:02:23,950 --> 00:02:32,475 softs. The likes of cocoa and coffee trending to the downside. 43 00:02:32,475 --> 00:02:40,225 Energy markets starting to move up. And then elsewhere, obviously 44 00:02:40,225 --> 00:02:48,750 currencies probably a bit more choppy and US bonds obviously moving 45 00:02:48,750 --> 00:02:57,275 up, trending in the last while. But obviously it's been a 46 00:02:57,275 --> 00:03:06,575 good period, not just this month, last month, but for a few 47 00:03:06,575 --> 00:03:07,350 months. Mark,any thoughts on how performance has evolved more 48 00:02:37,635 --> 00:03:07,425 recently in the last few months? 49 00:03:07,830 --> 00:03:21,228 Well, what’s interesting is that, I like to say, after January 50 00:03:21,228 --> 00:03:38,280 we may have hit, at least in the metals and maybe some of the 51 00:03:38,280 --> 00:03:52,896 other markets, what I'll call peak bubble. And when you think about 52 00:03:52,896 --> 00:04:09,948 it, we had the spike in gold and silver. We've had a little bit 53 00:04:09,948 --> 00:04:12,384 of reversal. What'sreally crazy is, if you look at some of 54 00:03:14,480 --> 00:03:23,790 the softs, if you look at what happened to cocoa in the last year 55 00:03:23,790 --> 00:03:33,765 or so, we've gone from a 12 handle down to 4, which is the same 56 00:03:33,765 --> 00:03:41,745 with coffee. It's just been a tremendous move. And this is what 57 00:03:41,745 --> 00:03:50,390 we see repeatedly with a lot of bubbles. You could have a blow 58 00:03:50,390 --> 00:03:57,705 off top that might have some correction and then sometimes they'll 59 00:03:57,705 --> 00:04:06,350 just slowly grind lower. And this is where a lot of trend followers 60 00:04:06,350 --> 00:04:12,335 can make money on both sides of the market. Marketgoes higher, 61 00:03:13,065 --> 00:03:34,005 you take on a little of that speculative fever on the upside, 62 00:03:34,005 --> 00:04:00,180 then, if you can learn to get out at near the tops, then you can 63 00:04:00,180 --> 00:04:12,395 also play it on the other side. 64 00:04:13,300 --> 00:04:20,483 Absolutely, yeah. And that's certainly the case, as you say, with 65 00:04:20,483 --> 00:04:28,319 cocoa, coffee, and obviously crude oil in that big move up back 66 00:04:28,319 --> 00:04:36,155 in the 2000s as well, was another example of those kind of 67 00:04:36,155 --> 00:04:42,032 extreme moves. But yeah, I mean, it's interesting mentioning 68 00:04:42,032 --> 00:04:49,215 bubble. And I know in your notes, in preparation for today, 69 00:04:49,215 --> 00:04:52,510 you're talking about peak bubble. Imean,coming into here 70 00:04:16,565 --> 00:04:25,054 and over the last few months, there's been a lot of talk about 71 00:04:25,054 --> 00:04:33,543 bubbles in the market and tech stock, etc. And obviously a lot of 72 00:04:33,543 --> 00:04:42,032 the momentum has already come out of that. So, is this a pause, 73 00:04:42,032 --> 00:04:51,174 do you think, or are we in the midst of the initial deflating, do 74 00:04:51,174 --> 00:04:52,480 you think? 75 00:04:53,280 --> 00:05:32,298 Well, it's hard to say because you have to go back to what causes 76 00:05:32,298 --> 00:06:05,742 bubbles in the first place. So, you're going to need the fuel, 77 00:06:05,742 --> 00:06:41,973 which is excess money, which I think we still have. But you also 78 00:06:41,973 --> 00:07:09,843 need a narrative and a story associated with a bubble. Whatwe 79 00:04:57,020 --> 00:05:23,200 find with a lot of bubbles is that it's for those assets that are 80 00:05:23,200 --> 00:05:47,510 very hard to value. Because if they're hard to value then you could 81 00:05:47,510 --> 00:06:09,950 be able to get stories or narratives that lead to extreme price 82 00:06:09,950 --> 00:06:28,650 moves. Now, an interesting piece of research, studying, it was 83 00:06:28,650 --> 00:06:45,480 talked about the retail habitat. So, analysts have looked 84 00:06:45,480 --> 00:07:09,790 at retail investors and sort of said what seems to be their characteristics? 85 00:04:53,280 --> 00:05:11,010 Oneof their characteristics is that they actually gravitate to 86 00:05:11,010 --> 00:05:26,967 hard-to-value stocks. So, stocks and other assets that are 87 00:05:26,967 --> 00:05:46,470 hard to value, if they have an optimistic narrative, they could 88 00:05:46,470 --> 00:06:11,292 lead to sort of extremes. So, that's what we sort of see with a 89 00:06:11,292 --> 00:06:34,341 lot of tech stocks. If you don't know how to discount the future 90 00:06:34,341 --> 00:06:55,617 cash flows from these firms, then it's harder to value. You could 91 00:06:55,617 --> 00:07:09,801 be extremely optimistic and that leads to bubbles. Inthe commodities 92 00:04:59,133 --> 00:05:20,594 markets I think we are in a different situation. People sometimes 93 00:05:20,594 --> 00:05:42,055 forget that with a lot of commodities, especially if it's not 94 00:05:42,055 --> 00:06:01,565 an annual crop, there supply constraints; mining for example - 95 00:06:01,565 --> 00:06:21,075 supply constraints. So, if you have some optimistic euphoria, whether 96 00:06:21,075 --> 00:06:48,389 it be in silver or gold, where are you going to get the supply? 97 00:06:48,389 --> 00:07:09,850 The supply is constrained, so, you could have these demand shocks. 98 00:04:53,280 --> 00:05:21,048 Andthat's what we've seen with the gold market and the silver 99 00:05:21,048 --> 00:05:51,130 market is that we're not going to see new supply. In fact, we'll 100 00:05:51,130 --> 00:06:21,212 sort of say that some of the discount in gold, from the high, 101 00:06:21,212 --> 00:06:48,980 is just related to the fact that there are fewer retail investors 102 00:06:48,980 --> 00:07:09,806 right now because of the Chinese lunar New Year. 103 00:07:10,950 --> 00:07:14,400 Okay, interesting. Well as you say, gravitating to the hard-to-value 104 00:07:14,400 --> 00:07:17,850 market, it certainly makes sense; the tech stocks, crypto and 105 00:07:17,850 --> 00:07:21,645 obviously there's been a lot of retail participation in the gold 106 00:07:21,645 --> 00:07:26,130 market of late. I was looking at some data there from the World 107 00:07:26,130 --> 00:07:28,890 Gold Council and they were highlighting, obviously, central 108 00:07:28,890 --> 00:07:33,030 bank buying has been an important theme but it was less in 109 00:07:33,030 --> 00:07:36,825 2025 than in 2024, and certainly ETF flow was very much 110 00:07:36,825 --> 00:07:40,965 the dominant driver in the latter part of the year. So, that's 111 00:07:40,965 --> 00:07:44,415 very much consistent with, you know, much more retail participation. 112 00:07:44,415 --> 00:07:47,520 Obviously, as prices rise, that encourages new entrants into 113 00:07:47,520 --> 00:07:50,625 the market. So, I think that's definitely fair there. 114 00:07:51,120 --> 00:08:03,671 Well, the central bank issue is very central to gold markets. 115 00:08:03,671 --> 00:08:18,504 We'll talk about this a little bit later when we talk about the 116 00:08:18,504 --> 00:08:32,196 US dollar, and US assets as a safe asset. Those countries that 117 00:08:32,196 --> 00:08:45,888 feel as though there might be a potential risk of sanctions have 118 00:08:45,888 --> 00:08:57,298 been bigger buyers. Their central banks have been bigger buyers 119 00:08:57,298 --> 00:09:01,862 than other central banks. So,we'll say, China has been a big 120 00:08:01,224 --> 00:08:16,380 central bank buyer as well as some of the other countries, that 121 00:08:16,380 --> 00:08:30,273 we know, that are worried about sanctions from the US, have 122 00:08:30,273 --> 00:08:42,903 been bigger central bank buyers. Some are on the periphery 123 00:08:42,903 --> 00:08:59,322 of where there could be war; like Poland has been a big buyer 124 00:08:59,322 --> 00:09:01,848 of gold. Andso, we'll sort of see that this is serving as a quasi-safe 125 00:08:04,250 --> 00:08:16,370 asset. So, if you have foreign reserves or you have your excess 126 00:08:16,370 --> 00:08:28,490 reserves, if you're a central bank says, like, well if I don't 127 00:08:28,490 --> 00:08:42,630 want to put it in dollars, where else can I put my money? And 128 00:08:42,630 --> 00:08:54,750 they say, well, if I diversify, I should put some in gold. 129 00:08:54,750 --> 00:09:01,820 This is what we've seen has changed. 130 00:09:02,040 --> 00:09:21,995 Yeah, for sure. And you're right, I mean the data I looked at, 131 00:09:21,995 --> 00:09:41,950 I think it had the National Bank of Poland that was the largest 132 00:09:41,950 --> 00:10:01,905 buyer in that period, and also a lot of the former Soviet countries 133 00:10:01,905 --> 00:10:20,325 ending in ‘stan’. Places that are hard to pronounce, not just Uzbekistan 134 00:10:20,325 --> 00:10:35,675 but other places like that. So, that's certainly consistent with 135 00:10:35,675 --> 00:10:40,280 what you're saying. Now,interestingly you touched on 136 00:09:10,655 --> 00:09:27,885 narrative there, which I think is quite interesting and obviously 137 00:09:27,885 --> 00:09:48,561 something that we're seeing in markets at the moment is a shift 138 00:09:48,561 --> 00:10:10,960 in the narrative around AI. If we went back 12 months, maybe, AI 139 00:10:10,960 --> 00:10:29,913 was seen as a positive influence, it was going to change 140 00:10:29,913 --> 00:10:40,251 the world. It was all positive. Morespending was a good 141 00:09:07,100 --> 00:09:21,268 thing, it was going to benefit the likes of Nvidia, and all of those 142 00:09:21,268 --> 00:09:28,352 chips and infrastructure providers. Obviously, that's shifted 143 00:09:28,352 --> 00:09:40,496 as we've come into this year. We've had the SaaS apocalypse (I 144 00:09:40,496 --> 00:09:50,616 struggled to get that one out), the software sell-off, and, 145 00:09:50,616 --> 00:10:00,736 I suppose, more generally, that shift in the narrative that 146 00:10:00,736 --> 00:10:11,868 clearly there will be winners and losers from this AI revolution, 147 00:10:11,868 --> 00:10:21,988 that's one thing. And then secondly, obviously, in terms of 148 00:10:21,988 --> 00:10:32,108 market reaction to earnings numbers and you know, the releases 149 00:10:32,108 --> 00:10:40,204 about the amount of Capex spending around AI. So,just curious, 150 00:09:09,228 --> 00:09:35,584 your take on that shifting narrative. I mean, what does that 151 00:09:35,584 --> 00:10:06,732 suggest to you, in terms of… Is that symptomatic more of the end 152 00:10:06,732 --> 00:10:30,692 of this move or just a different transition, a different 153 00:10:30,692 --> 00:10:40,276 phase of the move? 154 00:10:41,020 --> 00:11:03,263 Well, let's talk about what I mean by the word narrative. So, when 155 00:11:03,263 --> 00:11:22,084 we say the word narrative, oftentimes we say, well we think 156 00:11:22,084 --> 00:11:44,327 of it as a story, which is true. But I think that narrative 157 00:11:44,327 --> 00:12:04,859 is the use of storytelling when we don't have what we'll call 158 00:12:04,859 --> 00:12:21,969 countable risk. So, let's say that there's a difference between 159 00:12:21,969 --> 00:12:39,079 risk and uncertainty. So, a risk is something that's measurable. 160 00:12:39,079 --> 00:12:56,189 I can count it. So that would be our volatility. So,if you say, 161 00:10:51,420 --> 00:11:20,020 well, what's the volatility market? You say, let's look at the 162 00:11:20,020 --> 00:11:43,420 VIX, okay, it's something that's countable even though that's 163 00:11:43,420 --> 00:12:09,420 a market expectation from the options market. But then there's 164 00:12:09,420 --> 00:12:38,020 uncertainty or that which is not countable because I don't have 165 00:12:38,020 --> 00:12:56,220 any past events I could look to. So,when I think of AI and the 166 00:11:05,604 --> 00:11:42,480 AI revolution, since this is new technology, I don't have a way 167 00:11:42,480 --> 00:12:10,137 to actually measure the countable risk. So, therefore I'm 168 00:12:10,137 --> 00:12:40,867 in the realm of uncountable, which is uncertainty. And therefore, 169 00:12:40,867 --> 00:12:56,232 I have to use narrative. Andthen when I use the narrative, 170 00:10:49,700 --> 00:11:07,060 I have to come up with a metaphor or story for what I think 171 00:11:07,060 --> 00:11:20,700 might happen in the future. And in this particular case they, 172 00:11:20,700 --> 00:11:36,820 say, will I have to discount or sort of say the potential use 173 00:11:36,820 --> 00:11:54,180 of AI in the future. And we'll sort of say that AI is revolutionizing 174 00:11:54,180 --> 00:12:10,300 a lot of our behavior. It's going to revolutionize a lot of businesses. 175 00:12:10,300 --> 00:12:22,700 It's going to increase productivity. But is the growth going 176 00:12:22,700 --> 00:12:37,580 to be what is embedded right now in expectations? Is it going 177 00:12:37,580 --> 00:12:56,180 to be a little bit lower? Is it going to be a little bit higher? 178 00:10:41,020 --> 00:11:17,896 Andbecause we can't sort of count that, that's where the narrative 179 00:11:17,896 --> 00:11:54,772 issue comes in, you could have a very optimistic narrative and then 180 00:11:54,772 --> 00:12:31,648 discount the price with those values or it may be something in 181 00:12:31,648 --> 00:12:56,232 between a failure and these overly optimistic forecasts. 182 00:12:56,260 --> 00:12:59,516 We’ve kind of fallen into more of rangy markets, particularly with 183 00:12:59,516 --> 00:13:03,364 respect to equities in the last while. I mean, if you look at 184 00:13:03,364 --> 00:13:06,620 performance, say, over the last four weeks or so, bonds have 185 00:13:06,620 --> 00:13:09,876 actually gone up, equities have come down a little bit. You 186 00:13:09,876 --> 00:13:14,020 had the kind of the escalation of the run up in metals and then 187 00:13:14,020 --> 00:13:16,980 a correction. Currencies have been ranging, so nothing… I suppose, 188 00:13:16,980 --> 00:13:20,236 underneath, maybe the big theme has been the rotation in markets. 189 00:13:20,236 --> 00:13:22,012 Is that probably fair to say? 190 00:13:22,350 --> 00:13:38,694 Absolutely. I think this is where, when we talk about what are 191 00:13:38,694 --> 00:13:53,676 the major themes in the marketplace, we'll sort of say that 192 00:13:53,676 --> 00:14:08,658 rotation is probably the major theme for 2026. So, the rotation 193 00:14:08,658 --> 00:14:25,002 has been out of the US into European and global stocks. It's 194 00:14:25,002 --> 00:14:35,898 been out of US bonds, to some degree… Now,when we say out of bonds 195 00:13:29,917 --> 00:13:45,051 we want to be precise. There's been a lot of buying of US bonds, 196 00:13:45,051 --> 00:13:56,942 even by foreign investors. But, at this particular time, a lot 197 00:13:56,942 --> 00:14:09,914 of them are now hedging those bonds as opposed to unhedged. And 198 00:14:09,914 --> 00:14:23,967 we see US pension funds flows moving from the US and to outside 199 00:14:23,967 --> 00:14:35,858 the United States. And so that's where the rotation comes in. 200 00:13:22,350 --> 00:13:35,328 We'reseeing the difference in, say, software companies have 201 00:13:35,328 --> 00:13:51,190 fallen, more real asset companies have gone up. The AI stocks 202 00:13:51,190 --> 00:14:08,494 have fallen also. And then we'll sort of say that there's been 203 00:14:08,494 --> 00:14:25,798 more buying across a diverse set of companies than what we saw 204 00:14:25,798 --> 00:14:35,892 in 2025, where everything was more concentrated. 205 00:14:37,030 --> 00:14:43,113 And I mean, one of the interesting features has been, if 206 00:14:43,113 --> 00:14:50,302 you look at it from more of a sectoral perspective, you know, you've 207 00:14:50,302 --> 00:14:55,832 seen maybe industrials and materials (which would be kind of 208 00:14:55,832 --> 00:15:01,915 typically seen as cyclical sectors) doing well, and at the same 209 00:15:01,915 --> 00:15:07,445 time consumer staples doing well (which would be typically more 210 00:15:07,445 --> 00:15:14,634 kind of defensive stock). So, I mean, from that lens, a bit of 211 00:15:14,634 --> 00:15:21,270 an inconsistency, I guess we are seeing a rotation out of growth 212 00:15:21,270 --> 00:15:22,376 into value. ButI mean, is there, I suppose, a coherent story 213 00:14:46,480 --> 00:14:56,875 around that narrative apart from… Okay, I hear what you're saying 214 00:14:56,875 --> 00:15:09,160 about add a view, as it were, which is underpinned by a fairly 215 00:15:09,160 --> 00:15:19,555 obvious kind of story. But the other rotations, what's driving those 216 00:15:19,555 --> 00:15:22,390 do you think? 217 00:15:22,400 --> 00:15:29,045 Well, let's go back to some of the facts that we sort of see. One 218 00:15:29,045 --> 00:15:34,361 (and I'll be leading the witness) is, what do you think of 219 00:15:34,361 --> 00:15:38,348 the US Inflation relative to emerging market inflation? Which 220 00:15:38,348 --> 00:15:42,335 do you think is higher or lower right now? 221 00:15:43,810 --> 00:15:48,166 Well, you wouldn't guess US, but I'm guessing this is a trick 222 00:15:48,166 --> 00:15:48,529 question. 223 00:15:50,450 --> 00:16:03,372 It's sort of a trick question, but I found this very amazing that 224 00:16:03,372 --> 00:16:18,282 if you look at a basket of EM countries and then you look at their 225 00:16:18,282 --> 00:16:28,222 inflation, it's actually lower than US Inflation, even now. So, 226 00:16:28,222 --> 00:16:38,162 obviously US Inflation was peaking after the pandemic. But we'll 227 00:16:38,162 --> 00:16:52,078 sort of say that, even now, US inflation is at around 3%. Our core 228 00:16:52,078 --> 00:17:03,012 inflation is higher than what we're seeing in emerging market countries. 229 00:15:50,450 --> 00:16:22,202 So,this is a really big theme because what that means is that EM 230 00:16:22,202 --> 00:16:49,418 is much more attractive than the US just from the inflation story. 231 00:16:49,418 --> 00:17:03,026 So, that's going to cause rotation. So,if you see that kind 232 00:15:56,330 --> 00:16:08,090 of behavior going on, then you have to say, well, what's really 233 00:16:08,090 --> 00:16:22,790 going on? And I think that the major theme, that I see, is two factors. 234 00:16:22,790 --> 00:16:34,550 We have this K shaped economy where we have real differences between 235 00:16:34,550 --> 00:16:47,290 the real economy and then the asset economy, in the US, and that's 236 00:16:47,290 --> 00:16:58,070 being played out in the behavior of assets such as stocks 237 00:16:58,070 --> 00:17:02,970 and bonds around the world. 238 00:17:05,350 --> 00:17:13,738 And, I mean, you touch on the EM inflation versus US inflation. 239 00:17:13,738 --> 00:17:24,223 And I guess one of the upshots of that is that real yields, in EM, 240 00:17:24,223 --> 00:17:33,310 are higher and more attractive than in the US. I mean, I suppose, 241 00:17:33,310 --> 00:17:41,698 linked to this is, I suppose, historically the EM had a premium 242 00:17:41,698 --> 00:17:47,989 because policy was less credible, central banks were less 243 00:17:47,989 --> 00:17:54,280 independent, etc. They weren't as rigidly pursuing inflation targeting. 244 00:17:05,350 --> 00:17:19,340 Butnow everything is shifting. They're kind of more adhering 245 00:17:19,340 --> 00:17:31,931 to those inflation targets, whereas credibility is maybe diminishing 246 00:17:31,931 --> 00:17:48,719 in the US. So, I suppose is that relative attractiveness of the 247 00:17:48,719 --> 00:17:54,315 US versus EM shifting? 248 00:17:55,951 --> 00:18:04,789 Yes, I think the attractiveness is shifting. And we'll 249 00:18:04,789 --> 00:18:19,519 just say that part of this has to do with the major story of uncertainty. 250 00:18:19,519 --> 00:18:29,339 Now, I follow, pretty closely, the trade uncertainty indices; geopolitical 251 00:18:29,339 --> 00:18:37,195 uncertainty, monetary uncertainty. Some of those have peaked 252 00:18:37,195 --> 00:18:50,943 in the fall and some of them have come down. But let's say, trade 253 00:18:50,943 --> 00:18:59,781 uncertainty is still at very high levels. Overall uncertainty 254 00:18:59,781 --> 00:19:08,619 is high, although, again, it's come off its peak. Butwhen there's 255 00:18:01,144 --> 00:18:20,185 more uncertainty in, let's say, a G7 country or, in particular, 256 00:18:20,185 --> 00:18:40,957 the US, that's going to have an impact on whether foreign investors 257 00:18:40,957 --> 00:19:05,191 or even US investors look at US as a safe asset. So, that's on 258 00:19:05,191 --> 00:19:08,653 one level. Andalso, if there's more uncertainty, well, I'm 259 00:18:05,807 --> 00:18:23,055 going to need a premium to hold a risky asset in a highly uncertain 260 00:18:23,055 --> 00:18:34,143 country versus a less uncertain country. And so, we're 261 00:18:34,143 --> 00:18:47,695 also seeing sort of flows going out there because you say, 262 00:18:47,695 --> 00:19:01,247 well, maybe I could discount those cash flows better in other 263 00:19:01,247 --> 00:19:08,639 countries than the US right now. 264 00:19:08,671 --> 00:19:23,553 Just on this EM issue, I mean, is this a structural change, do you 265 00:19:23,553 --> 00:19:39,498 think, or we just had a point in time where this is a sweet spot 266 00:19:39,498 --> 00:19:53,317 for EM? I'm kind of thinking in terms of the overall regime that 267 00:19:53,317 --> 00:19:58,632 we're in at the moment. Obviouslyit's a changed economic 268 00:19:11,646 --> 00:19:19,381 regime from what we were in in the 2010s, which was secular stagnation. 269 00:19:19,381 --> 00:19:25,331 Obviously, we're in a higher nominal GDP environment, more inflation 270 00:19:25,331 --> 00:19:30,686 uncertainty, there are various features, from an economic perspective. 271 00:19:30,686 --> 00:19:36,636 But we're also seeing this shift in the relative attractiveness 272 00:19:36,636 --> 00:19:43,776 of EM debt versus a developed or, particularly, US debt. But I 273 00:19:43,776 --> 00:19:50,916 mean, could EM sovereigns be seen as the new safe assets, really? 274 00:19:50,916 --> 00:19:58,651 Or once we get to the next crisis, will all of this unravel? 275 00:19:59,131 --> 00:20:13,975 Well, I'm spending more time thinking about the concept of a safe 276 00:20:13,975 --> 00:20:28,819 asset as a relative concept as opposed to an absolute concept. I 277 00:20:28,819 --> 00:20:44,900 think a lot of people, when they think about a safe asset, well 278 00:20:44,900 --> 00:20:59,744 they, they've always viewed it as the US dollar or treasuries as 279 00:20:59,744 --> 00:21:09,640 a safe asset, and it was as absolute. And I think now we have 280 00:20:05,263 --> 00:20:16,505 to think of safety as a relative concept. You could still 281 00:20:16,505 --> 00:20:29,791 be a safe asset for US treasury, but it might be less safe 282 00:20:29,791 --> 00:20:43,077 than it was a year ago or two years ago. So consequently, your 283 00:20:43,077 --> 00:20:54,319 relative safety has gone down even though, on an absolute basis, 284 00:20:54,319 --> 00:21:05,561 you're safe. That means that people are going to diversify more 285 00:21:05,561 --> 00:21:09,649 of their asset flows. Havewe seen a fundamental shift in EM? That's 286 00:20:12,928 --> 00:20:29,791 probably not my area of expertise and I think that those 287 00:20:29,791 --> 00:20:45,121 are longer-term structural changes. But we do see a situation 288 00:20:45,121 --> 00:21:03,517 in that relative safety, or the view of the relative safety of 289 00:21:03,517 --> 00:21:09,649 US assets has fallen. 290 00:21:10,311 --> 00:21:14,307 Yeah, fair enough. I mean it is something I've been thinking about. 291 00:21:14,307 --> 00:21:17,970 I mean, you do hear this comment, sometimes, that bonds are 292 00:21:17,970 --> 00:21:21,300 not going to diversify equities anymore, look at the bond 293 00:21:21,300 --> 00:21:24,963 equity correlation. But it's very definitive when you say it like 294 00:21:24,963 --> 00:21:28,959 that, whereas I don't think that's the case. I mean, you can 295 00:21:28,959 --> 00:21:31,956 certainly envisage certain scenarios where we have crises, the 296 00:21:31,956 --> 00:21:35,286 Fed cuts rate aggressively, and bonds rally, while equities go 297 00:21:35,286 --> 00:21:39,615 down, and bonds do diversify. So, you know, I would think in more, 298 00:21:39,615 --> 00:21:44,277 as you say, like kind of less safe or less reliable as a diversifier, 299 00:21:44,277 --> 00:21:48,939 but it’s not yes or no. It's more, less so than in the past. 300 00:21:49,691 --> 00:22:12,131 Well, this is the key overall theme. You know, obviously, I'm a 301 00:22:12,131 --> 00:22:30,831 trend follower, a strong believer in trend follower, but also 302 00:22:30,831 --> 00:22:57,011 a global macro guy. And the reason why I am a global macro person 303 00:22:57,011 --> 00:23:17,581 is because I think we go through regime changes, or there 304 00:23:17,581 --> 00:23:38,151 are different regimes that we might be in where relationships change 305 00:23:38,151 --> 00:24:00,591 through time. By being aware of (we'll call it) the regime changes, 306 00:24:00,591 --> 00:24:24,901 or how all of our analysis is conditional, then you might be able 307 00:24:24,901 --> 00:24:49,211 to get an edge on how you should form your portfolios or where 308 00:24:49,211 --> 00:25:00,431 you think the opportunities might exist. Now,I was going to 309 00:22:05,861 --> 00:22:44,669 start with a quote earlier in our podcast. That was from Peter 310 00:22:44,669 --> 00:23:23,477 Lynch from Fidelity, and he has worked for Fidelity for a number 311 00:23:23,477 --> 00:24:08,753 of years. I don't believe this quote, but I thought it was a great 312 00:24:08,753 --> 00:24:54,029 one. He said, ‘if you spend 14 minutes on macro then that's 12 minutes 313 00:24:54,029 --> 00:25:00,497 too long’. So,I don't know if I will actually believe that. I think 314 00:22:15,131 --> 00:22:40,571 you should use the whole 14 minutes, or maybe more, to spend 315 00:22:40,571 --> 00:23:08,131 time on macro. But I think that there is this view that you 316 00:23:08,131 --> 00:23:33,571 don't need to know what's going on in the global macro economy. 317 00:23:33,571 --> 00:24:01,131 And I'll take the opposite point of view in that we do need 318 00:24:01,131 --> 00:24:30,811 to do this. That has a big impact on all types of assets because 319 00:24:30,811 --> 00:24:49,891 most of our asset price relationships are conditional in 320 00:24:49,891 --> 00:25:00,491 the environment we live in. Let'sgive a simple example. There 321 00:22:05,813 --> 00:22:38,057 is an interesting paper I was looking at, talking about stock beta. 322 00:22:38,057 --> 00:23:15,675 And if you just look at what the beta is for an individual stock 323 00:23:15,675 --> 00:23:53,293 or what the beta is for a given asset, and once you break it 324 00:23:53,293 --> 00:24:17,476 down into different regimes, whether we're in crisis, non-crisis, 325 00:24:17,476 --> 00:24:44,346 recession, non-recession, the concept of a beta for an individual 326 00:24:44,346 --> 00:25:00,531 stock is a fairly fluid one. Irememberone discussion I was having 327 00:22:02,410 --> 00:22:27,848 when I was at John Henry, he said like what's the beta for the 328 00:22:27,848 --> 00:22:51,469 stock? And I said, well, it's this, but if we measure it differently, 329 00:22:51,469 --> 00:23:15,090 we'll get a different beta. He goes, well what do you mean there's 330 00:23:15,090 --> 00:23:36,894 a different beta? He said there should be one. What is the 331 00:23:36,894 --> 00:24:00,515 beta? And I said, well, there isn't a beta, there are many betas. 332 00:24:00,515 --> 00:24:24,136 His hands were thrown up in frustration and he said, I ask you 333 00:24:24,136 --> 00:24:45,940 a simple question, can't you give me a simple answer? And the 334 00:24:45,940 --> 00:25:00,476 answer is no, there are no simple answers. Nowthat could just 335 00:22:03,319 --> 00:22:44,203 be because I'm an economist and there is always, ‘on the other 336 00:22:44,203 --> 00:23:28,494 hand’. In some sense, a lot of our relationships that we see are 337 00:23:28,494 --> 00:24:12,785 conditional and that's why we have to spend a lot of time looking 338 00:24:12,785 --> 00:24:40,041 at different regimes, looking at different environments because 339 00:24:40,041 --> 00:25:00,483 that's how we create an edge. 340 00:25:01,331 --> 00:25:22,343 Yeah, fair enough. Well, I agree with you. I mean Peter Lynch 341 00:25:22,343 --> 00:25:46,857 was a bottom-up stock picker. So, I mean in terms of how he ran 342 00:25:46,857 --> 00:26:02,616 money, obviously hugely successfully, he wasn't a macro oriented 343 00:26:02,616 --> 00:26:25,379 investor. But I think for the rest of us, if we're thinking about 344 00:26:25,379 --> 00:26:41,138 portfolio construction, it’s certainly good to be cognizant of 345 00:26:41,138 --> 00:26:53,395 the macro picture and the regime changes. Butjust on the regime 346 00:25:11,516 --> 00:25:35,960 change I mean, we always talk about it, you know, that there's 347 00:25:35,960 --> 00:26:00,404 a regime change underway. And I think you can certainly paint that 348 00:26:00,404 --> 00:26:20,774 picture, as I say, with secular stagnation in the 2010s, 349 00:26:20,774 --> 00:26:37,070 deglobalization, all of that falling inflation, inflation targeting, 350 00:26:37,070 --> 00:26:53,366 etc., all of these things sort of changed. Now.you could look at 351 00:25:12,316 --> 00:25:34,286 the equity market and you wouldn't really see much evidence 352 00:25:34,286 --> 00:26:05,044 of a regime shift. It went up back then, it's still going up now. 353 00:26:05,044 --> 00:26:33,605 So, nothing there. But it is curious to think about what are the 354 00:26:33,605 --> 00:26:53,378 elements where we're seeing that change in the regime. Andso, 355 00:25:03,907 --> 00:25:16,787 the bond/equity correlation is obviously one example of that. And 356 00:25:16,787 --> 00:25:34,819 that was very much in 2022. I guess another one is maybe the US 357 00:25:34,819 --> 00:25:47,699 dollar. Going prior to Liberation Day, the dollar was kind 358 00:25:47,699 --> 00:26:04,443 of strong on talk of tariffs and it was rising on relative interest 359 00:26:04,443 --> 00:26:17,323 rate differentials. And since then that relationship has been weaker. 360 00:26:17,323 --> 00:26:31,491 And obviously you could say there's some kind of negative premium 361 00:26:31,491 --> 00:26:45,659 or a discount or whatever being applied to the dollar, maybe 362 00:26:45,659 --> 00:26:53,387 on the basis of credibility, etc. Andmaybe gold is the flip side 363 00:25:12,076 --> 00:25:27,426 of that. Historically you would have explained it in terms 364 00:25:27,426 --> 00:25:44,311 of real yields and the dollar, but obviously it's been accelerating 365 00:25:44,311 --> 00:26:02,731 in the last one to two years independent of those factors driven 366 00:26:02,731 --> 00:26:19,616 by central bank buying, ETF buying, etc. Anything else you look 367 00:26:19,616 --> 00:26:41,106 at or any other signs that you see, that point to, that that are 368 00:26:41,106 --> 00:26:53,386 evidence of, that regime change, do you think? 369 00:26:53,811 --> 00:27:37,829 Well, we can think of regime change on a number of different levels. 370 00:27:37,829 --> 00:28:18,461 So, you can think of regime changes and use sort of techniques 371 00:28:18,461 --> 00:28:55,707 called, like, hidden Markov processes to just sort of say like 372 00:28:55,707 --> 00:29:46,497 well, are we in a high vol / low vol regime, or risk-on / risk-off 373 00:29:46,497 --> 00:30:23,743 regime? And I'm not saying that those are easier to forecast, 374 00:30:23,743 --> 00:30:57,603 but we could sort of say that's very data dependent. So,we 375 00:26:59,415 --> 00:27:38,643 do find, for example, if we're in a higher risk regime or a higher 376 00:27:38,643 --> 00:27:58,257 risk environment, that the behavior difference, strategies, 377 00:27:58,257 --> 00:28:29,079 and tactics, are different than if you're in a low volatility 378 00:28:29,079 --> 00:29:02,703 regime. And in particular, let's look at the VIX. What you find 379 00:29:02,703 --> 00:29:39,129 is that the VIX is a very skewed distribution. We do find that, 380 00:29:39,129 --> 00:30:15,555 once it gets above like 25 and it's moving higher (so it's above 381 00:30:15,555 --> 00:30:46,377 a certain threshold level and moving higher), it's bad news for 382 00:30:46,377 --> 00:30:57,585 a lot of assets. Sothen if it gets to above you know, around 40 383 00:27:34,445 --> 00:28:18,773 whatever, then we know it's sort of peaking, and then we know 384 00:28:18,773 --> 00:29:06,795 it's going to come back the other way, and that's good news to 385 00:29:06,795 --> 00:29:43,735 get back in. So, there's a nice nonlinear relationship, and 386 00:29:43,735 --> 00:30:28,063 we could use sort of like techniques to look at high versus 387 00:30:28,063 --> 00:30:57,615 low vol environments. So, we could do that. Thereare also regimes 388 00:27:07,547 --> 00:27:41,887 when we think about the environment, you know, what type 389 00:27:41,887 --> 00:28:23,095 of monetary environment we are in and what type of fiscal environment 390 00:28:23,095 --> 00:29:04,303 we’re in. We could think about, you know, the view of geopolitical, 391 00:29:04,303 --> 00:29:38,643 if there's high level of uncertainty, that's going to change 392 00:29:38,643 --> 00:30:19,851 the demand for safe assets. So, there's, we'll call it the price 393 00:30:19,851 --> 00:30:47,323 driven regimes. There are policy structural regimes. There 394 00:30:47,323 --> 00:30:57,625 are sentiment regimes. So,we could decompose these different types 395 00:27:19,284 --> 00:28:06,591 of impacts. And so, in a simple podcast we can be flip with 396 00:28:06,591 --> 00:28:42,981 our definitions but when you look underneath the surface, you 397 00:28:42,981 --> 00:29:26,649 have to decompose this into a lot of different types of regimes 398 00:29:26,649 --> 00:30:21,234 and this is where you could be able to create your edge - so, by 399 00:30:21,234 --> 00:30:57,624 sort of decomposing the kind of regimes that you're in. Nowwe 400 00:27:03,011 --> 00:27:44,411 do know, for example, the stock equity correlation regime, 401 00:27:44,411 --> 00:28:30,411 which is usually negative, inflation goes up, usually then that 402 00:28:30,411 --> 00:29:21,011 stock/bond correlation is also going to go higher. So, it's inflation 403 00:29:21,011 --> 00:30:16,211 dependent. So, you can use sort of third factors to give us 404 00:30:16,211 --> 00:30:57,611 some indication of what kind of environment we're in. Andwhy 405 00:26:58,051 --> 00:27:27,731 do we want to look at that? Because if, let's say, that the stock/bond 406 00:27:27,731 --> 00:27:53,171 correlation changes, well that has a big impact on what type of 407 00:27:53,171 --> 00:28:14,371 alternatives you want to buy. If the stock/bond correlation is 408 00:28:14,371 --> 00:28:35,571 negative, and highly negative, and rates are reasonably high… Let 409 00:28:35,571 --> 00:29:01,011 me put this way, I don't really need to buy alternative strategies. 410 00:29:01,011 --> 00:29:24,331 But when that correlation starts to go positive and rates are 411 00:29:24,331 --> 00:29:54,011 lower, they say, well, that's a great time to own a lot of hedge 412 00:29:54,011 --> 00:30:17,331 fund strategies or other types of strategies. So, that's a perfect 413 00:30:17,331 --> 00:30:42,771 example where we could use sort of regime analysis to help us 414 00:30:42,771 --> 00:30:57,611 build portfolios, how we look at them. Andthen from a, we'll say, 415 00:27:15,645 --> 00:27:59,313 more micro basis, we know that when we build portfolios, even based 416 00:27:59,313 --> 00:28:35,703 on trends, the correlation relationships may have a big impact 417 00:28:35,703 --> 00:29:23,010 on the kind of risk exposures we have. So, knowing the regime that 418 00:29:23,010 --> 00:30:10,317 we're in could help us, on the margin, change our tilts and our 419 00:30:10,317 --> 00:30:50,346 exposures to make sure that we're not overexposed to one sector 420 00:30:50,346 --> 00:30:57,624 versus another. 421 00:31:03,971 --> 00:31:14,275 One of the topics you touched on there was the kind of the monetary 422 00:31:14,275 --> 00:31:22,371 regime, and I think you might have mentioned the fiscal regime. 423 00:31:22,371 --> 00:31:31,939 But certainly, we're into an era now where there is a lot more 424 00:31:31,939 --> 00:31:39,299 talk of possible fiscal dominance. I mean, people have been 425 00:31:39,299 --> 00:31:48,867 drawn around the term fiscal dominance for a while now as if we're 426 00:31:48,867 --> 00:31:50,341 here already. Idon'tthink we are quite there yet, but it is obviously 427 00:31:11,099 --> 00:31:17,633 a risk because fiscal dominance is when the debt and deficit 428 00:31:17,633 --> 00:31:24,167 levels are so great that they are dictating monetary policy. I 429 00:31:24,167 --> 00:31:30,701 know the pressure is on monetary policy at the moment, but 430 00:31:30,701 --> 00:31:37,829 we're not quite there yet. But taking that angle, that lens, what 431 00:31:37,829 --> 00:31:46,145 would you say about the regime now versus where we were from a kind 432 00:31:46,145 --> 00:31:50,303 of a fiscal and monetary dominance perspective? 433 00:31:50,581 --> 00:32:48,418 Right. Well, I think that this issue, or at least trying to frame 434 00:32:48,418 --> 00:33:41,806 a lot of the discussion about Fed independence. Frame who is going 435 00:33:41,806 --> 00:34:39,643 to be the chairman of the Fed. Frame how, so the tussle between, 436 00:34:39,643 --> 00:35:24,133 let's say the current US administration and the central bank 437 00:35:24,133 --> 00:36:17,521 has to be looked at through the idea of fiscal versus monetary 438 00:36:17,521 --> 00:36:21,970 dominance. Andso, of course there's a political agenda going 439 00:32:26,500 --> 00:33:18,383 on here, but I think sometimes we have to take ourselves away from 440 00:33:18,383 --> 00:34:10,266 the politics and look at the economics and try to say, what causes 441 00:34:10,266 --> 00:34:50,176 the tension between the Fed and the Treasury Department. So, 442 00:34:50,176 --> 00:35:34,077 we'll try to take out the administration. We'll call it Fed 443 00:35:34,077 --> 00:36:17,978 versus Treasury. And this is the issue of fiscal versus monetary 444 00:36:17,978 --> 00:36:21,969 dominance. Ina monetary dominant environment, we'll say the 445 00:32:36,781 --> 00:33:46,081 Fed could focus on, or the central bank can focus on inflation, 446 00:33:46,081 --> 00:34:43,831 employment, regulation, and the fiscal focus is just on financing 447 00:34:43,831 --> 00:36:04,681 - where can I minimize the cost of my debt? That's all I have 448 00:36:04,681 --> 00:36:22,006 to worry about. TheFed could do whatever they want. In fact, the 449 00:32:31,091 --> 00:33:15,652 fiscal side or the treasury could say, yeah, you could follow 450 00:33:15,652 --> 00:33:52,111 whatever you're doing because your objective is full employment. 451 00:33:52,111 --> 00:34:48,825 We're just going to then try to pick on the curve where we want 452 00:34:48,825 --> 00:35:37,437 to issue our debt and we're going to finance our deficits, which 453 00:35:37,437 --> 00:36:21,998 we don't think are going to be persistent, and everybody's happy. 454 00:31:50,581 --> 00:32:34,131 Okay.If we have to go into a fiscal dominant situation, well then, 455 00:32:34,131 --> 00:33:14,331 you know, the government and even the central bank, even if they're 456 00:33:14,331 --> 00:33:51,181 appointed for a long period of time, they're still authorized or 457 00:33:51,181 --> 00:34:24,681 there's oversight by Congress. Fiscal dominance says that debt is 458 00:34:24,681 --> 00:35:08,231 so high, we're going to put pressure on the central bank to use 459 00:35:08,231 --> 00:35:48,431 its powers to control interest rates and then buy up excess debt. 460 00:35:48,431 --> 00:36:21,931 So, like, well, is this abnormal? Should this be shocking? 461 00:31:50,581 --> 00:32:59,671 Well,look at World War II. The whole idea is that the Fed kept 462 00:32:59,671 --> 00:33:44,086 interest rates low because we're financing huge deficits to 463 00:33:44,086 --> 00:34:43,306 pay for the war. And then they are actively buying treasuries. So, 464 00:34:43,306 --> 00:35:52,396 this is what happened in World War II, post World War II. And then 465 00:35:52,396 --> 00:36:22,006 we had the Treasury Fed Accord. TheTreasury Fed Accord said 466 00:32:18,276 --> 00:33:24,744 that, well, now that we're going to split the Fed and Treasury 467 00:33:24,744 --> 00:34:31,212 working in tandem to lower the cost of financing because, in the 468 00:34:31,212 --> 00:35:26,602 effect of doing that, you raised inflation. Now we're going 469 00:35:26,602 --> 00:36:21,992 to allow them to sort of move in different directions. Now,what's 470 00:31:57,915 --> 00:32:38,252 important here is there had been another recent paper about debt. 471 00:32:38,252 --> 00:33:22,256 And then they talked about debt as a safe asset. Well, that's 472 00:33:22,256 --> 00:34:13,594 true except if we're having a war. If you buy debt, and if you're 473 00:34:13,594 --> 00:34:57,598 in a war scenario, because there's a lot of financing of wars 474 00:34:57,598 --> 00:35:30,601 and then there's usually inflation because of, perhaps, pent 475 00:35:30,601 --> 00:36:21,939 up demand after the war ends, you don't want to be a debt holder. 476 00:31:50,581 --> 00:32:48,331 Andwhat these authors actually said is that the pandemic 477 00:32:48,331 --> 00:33:51,856 had all the characteristics of a war. We had constrained demand, 478 00:33:51,856 --> 00:35:01,156 we had excess monetary policy, we sort of drove down interest rates, 479 00:35:01,156 --> 00:36:16,231 we had excess fiscal policy. It sort of offset the war on the 480 00:36:16,231 --> 00:36:22,006 pandemic. Sowhat happens when we came out of that, that's why we 481 00:32:29,353 --> 00:33:08,125 had a sort of inflation burst because of the excesses of fiscal 482 00:33:08,125 --> 00:33:43,666 policy. So, when you think about it, the excesses from that 483 00:33:43,666 --> 00:34:19,207 pandemic war is, now, we're still having that issue today because, 484 00:34:19,207 --> 00:35:01,210 in some sense, that we should have said, well, all of that spending 485 00:35:01,210 --> 00:35:39,982 was temporary, so we should have seen a big budget deficit increase. 486 00:35:39,982 --> 00:36:21,985 Now that the pandemic is over, we should have reversed all of that. 487 00:31:50,581 --> 00:33:51,205 Ifyou don't reverse that, well then you have a problem. And 488 00:33:51,205 --> 00:36:01,881 this is where now you have the tussle between the monetary and the 489 00:36:01,881 --> 00:36:21,985 fiscal side. 490 00:36:24,021 --> 00:36:41,703 A lot to get into there. Yes, as you say, you've got this tussle 491 00:36:41,703 --> 00:36:59,385 and as you rightly say, I mean Covid was treated just like a war. 492 00:36:59,385 --> 00:37:17,067 That was the kind of narrative at the time. And I think, in the 493 00:37:17,067 --> 00:37:32,223 UK, there was actually direct buying from the bank of England of 494 00:37:32,223 --> 00:37:47,379 bonds issued by the treasury, whereas in the US it wasn't direct 495 00:37:47,379 --> 00:38:02,535 buying, but in May, I mean, the treasury issued more debt and 496 00:38:02,535 --> 00:38:17,691 the Fed did more QE. And so, they didn't finance it directly, 497 00:38:17,691 --> 00:38:22,743 but effectively they did. Asyou say, we had a surge in deficits 498 00:36:37,548 --> 00:36:54,081 which haven't been addressed since then. I mean, looking at the 499 00:36:54,081 --> 00:37:10,614 precedent of, as you say, the 1940s, the justification, at the 500 00:37:10,614 --> 00:37:28,650 time, is obviously to try and keep debt financing costs down because 501 00:37:28,650 --> 00:37:48,189 you're in a war. It's also a form of financial repression of course 502 00:37:48,189 --> 00:38:03,219 as well, isn't it? Because, obviously, you pin down long-term 503 00:38:03,219 --> 00:38:19,752 rates and you allow inflation to be higher so bondholders experience 504 00:38:19,752 --> 00:38:22,758 negative returns. Andthen ultimately in that scenario, the 505 00:36:41,339 --> 00:37:08,553 Fed did flex its muscles again and demand its independence back. 506 00:37:08,553 --> 00:37:38,241 And that's why you had this kind of standoff between the Fed 507 00:37:38,241 --> 00:38:02,981 and the Treasury, which ultimately resulted in the Fed Treasury 508 00:38:02,981 --> 00:38:22,773 Accord, and the Fed got its independence back. So,this is all 509 00:36:30,269 --> 00:36:45,889 relevant in the current context because people are now saying 510 00:36:45,889 --> 00:37:03,071 you've got Kevin Warsh and you've got Scott Bessent, both linked 511 00:37:03,071 --> 00:37:15,567 via Stanley Druckenmiller, interestingly enough. But they obviously 512 00:37:15,567 --> 00:37:32,749 already have spoken about a possible new accord. But I mean, 513 00:37:32,749 --> 00:37:49,931 they’ve spoken about it in some kind of positive light. But 514 00:37:49,931 --> 00:38:07,113 I mean, you know, should we interpret that as more financial 515 00:38:07,113 --> 00:38:22,733 repression or what do you think that might look like? 516 00:38:23,031 --> 00:39:07,815 Well, the situation we're in right now, we'll sort of say that 517 00:39:07,815 --> 00:39:41,403 the current Treasury Department would love for the Fed 518 00:39:41,403 --> 00:40:22,455 to lower interest rates because that lowers the cost of debt 519 00:40:22,455 --> 00:41:03,507 and you could use that. Instead of paying off interest you 520 00:41:03,507 --> 00:41:52,023 can use that money for other purposes. It could be used for other 521 00:41:52,023 --> 00:41:55,755 expenditures. Soyou have the desire to do that and you sort of 522 00:38:52,371 --> 00:39:19,266 say, well, if you can't default on your debt, what's the 523 00:39:19,266 --> 00:39:48,606 easiest way to reduce the value of your nominal debt if you 524 00:39:48,606 --> 00:40:17,946 have inflation? So, you have a Treasury that would sort of say, 525 00:40:17,946 --> 00:40:44,841 I'd like to have lower interest rate to lower my financing 526 00:40:44,841 --> 00:41:16,626 costs. I really don't mind if, let's say there's a little bit higher 527 00:41:16,626 --> 00:41:41,076 inflation (even though they'll never say this because that reduces 528 00:41:41,076 --> 00:41:55,746 the real value of the debt). Andyou have, we'll call it, K shaped 529 00:38:49,623 --> 00:39:32,835 economy. And if you look at some of the data on the employment 530 00:39:32,835 --> 00:40:16,047 side, you can sort of say that I may, as a current administration, 531 00:40:16,047 --> 00:40:59,259 I may want to run an economy hotter because I don't believe the 532 00:40:59,259 --> 00:41:42,471 data that I have. And a perfect example we just had in January, 533 00:41:42,471 --> 00:41:55,767 the revisions for employment. Therevision for employment said 534 00:38:36,326 --> 00:39:02,916 that there was probably a million fewer jobs created because 535 00:39:02,916 --> 00:39:34,824 we changed the benchmark. The year before we actually did a seasonal 536 00:39:34,824 --> 00:40:04,073 adjustment and we got like lower job production. So, if you 537 00:40:04,073 --> 00:40:38,640 said, look, we didn't create a million jobs in the last year or 538 00:40:38,640 --> 00:41:15,866 so. The economy may not be doing as well as we expect, on the 539 00:41:15,866 --> 00:41:55,751 one leg of the K. So, you may want to, say, run my economy hotter. 540 00:38:23,031 --> 00:38:52,643 Now,you could also say that these benchmark revisions were caused 541 00:38:52,643 --> 00:39:24,947 by demographics. And so, let's say people are moving out of the 542 00:39:24,947 --> 00:39:49,175 United States (we'll call that euphemistically they're moving out 543 00:39:49,175 --> 00:40:21,479 of the United States), and you have less people in the workforce, 544 00:40:21,479 --> 00:40:59,167 well, then it may not matter as much. So, we don't see the unemployment 545 00:40:59,167 --> 00:41:28,779 number going higher. But you'd say there's this tussle between fiscal 546 00:41:28,779 --> 00:41:55,699 and monetary policy because we do have these competing interests. 547 00:38:23,031 --> 00:38:56,350 Nowthe interesting part and the reason how we get back to this 548 00:38:56,350 --> 00:39:21,980 recurring theme of always uncertainty and is that worse, he 549 00:39:21,980 --> 00:39:55,299 said is that we never really liked QE in the excesses of what 550 00:39:55,299 --> 00:40:28,618 we had. So, in some sense, if he actually follows through on his 551 00:40:28,618 --> 00:40:54,248 behaviors, he's going to reduce the amount of treasuries on 552 00:40:54,248 --> 00:41:19,878 the Fed balance sheet. That's reducing the overall liquidity. That's 553 00:41:19,878 --> 00:41:45,508 going to cause a major deleveraging which should have an 554 00:41:45,508 --> 00:41:55,760 impact on asset markets. Atthe same time, it's exactly what 555 00:38:43,149 --> 00:39:23,385 the Treasury does not want him to do. They like the fact that the 556 00:39:23,385 --> 00:40:03,621 Fed has a large balance sheet and we'll say like look at the size 557 00:40:03,621 --> 00:40:35,235 of the financing. And, again, Alan, I throw out these questions 558 00:40:35,235 --> 00:41:06,849 just to be provocative. You know, it's not the obvious answer. 559 00:41:06,849 --> 00:41:44,211 How much debt do you think the US actually issued, just last week? 560 00:41:44,211 --> 00:41:55,707 Just take a guess. 561 00:41:58,821 --> 00:42:00,141 Half a trillion or something there. 562 00:42:00,141 --> 00:42:15,002 So, you're close, but you're still off by about US$200 billion. 563 00:42:15,002 --> 00:42:31,214 In one week the treasury that we had sort of quarterly refinancing. 564 00:42:31,214 --> 00:42:50,128 Not all of this was new money. Some of this was rolled over. But, 565 00:42:50,128 --> 00:43:04,989 we did like a US$700 billion of treasuries being auctioned in 566 00:43:04,989 --> 00:43:19,850 one week. Now that was between bills, 2-years, 10-years. So, it 567 00:43:19,850 --> 00:43:26,605 was along the entire curve. Butthat's a huge amount of money. 568 00:42:08,912 --> 00:42:25,201 Now, the world can absorb that kind of stuff. But if, let's say 569 00:42:25,201 --> 00:42:38,984 that there was a change in people's buying habits for treasuries, 570 00:42:38,984 --> 00:42:55,273 that's a big number. And so, when you think about, in one week 571 00:42:55,273 --> 00:43:12,815 you could do that, then you look at the size of the balance sheet. 572 00:43:12,815 --> 00:43:26,598 That's a pretty good portion of the entire Fed balance sheet. 573 00:42:00,141 --> 00:42:20,406 Now,again, we have to look at what is the net new money and what 574 00:42:20,406 --> 00:42:35,267 is this total demand from a number of different sources. And 575 00:42:35,267 --> 00:42:52,830 you can cut sort of overnight repos. So, there's ways to do this, 576 00:42:52,830 --> 00:43:10,393 but that's a big number. And that should give people a sense of 577 00:43:10,393 --> 00:43:26,605 what we're dealing with in terms of the size of the issues. 578 00:43:27,491 --> 00:43:48,211 Yeah, well, as you say, the demand is there for it at the moment. 579 00:43:48,211 --> 00:44:07,451 I mean, you'd have to look to see where it's coming from. Obviously, 580 00:44:07,451 --> 00:44:22,251 that's shifted over time. Maybe it's less foreign. And also, 581 00:44:22,251 --> 00:44:38,531 they've talked about changing the SLR to make it more attractive 582 00:44:38,531 --> 00:44:51,891 for banks to hold it. So, there are things. Imean,coming 583 00:43:32,993 --> 00:43:55,001 back to Warsh, his idea is the smaller balance sheet but equally 584 00:43:55,001 --> 00:44:17,009 lower rates. And then, I guess, the upshot of all of that 585 00:44:17,009 --> 00:44:37,183 would be maybe more T-bill issuance to take advantage of lower 586 00:44:37,183 --> 00:44:51,855 short-term rates, if that's how it played out. Butone thing 587 00:43:33,377 --> 00:43:52,997 that we're missing in this… We're talking about fiscal dominance. 588 00:43:52,997 --> 00:44:10,655 Central banks are only targeting inflation. And now we're 589 00:44:10,655 --> 00:44:32,237 talking about adjusting rates to target, you know, to be cognizant 590 00:44:32,237 --> 00:44:51,857 of debt. But of course that forgets about asset prices. So,if 591 00:43:29,867 --> 00:43:44,123 we had this scenario of the Fed being more cognizant of debt 592 00:43:44,123 --> 00:43:56,003 considerations and lowering rates for that reason as well, not 593 00:43:56,003 --> 00:44:10,259 only could it be inflationary, but presumably, as you say, it would 594 00:44:10,259 --> 00:44:22,139 run the economy hot and potentially fuel asset bubbles as 595 00:44:22,139 --> 00:44:36,395 well. Which is another part of the discussion from the ‘90s, that 596 00:44:36,395 --> 00:44:51,839 kind of is kind of getting lost in the current debate, I think. 597 00:44:52,291 --> 00:45:39,065 Well, let's go back to what we started with. We always talk about 598 00:45:39,065 --> 00:46:25,839 regime changes. So, first this is that this could be a regime change 599 00:46:25,839 --> 00:47:12,613 if we get a new Fed chairman because he might have a different 600 00:47:12,613 --> 00:47:16,211 view. Nowlet me put this way, there could be a regime change because 601 00:45:19,799 --> 00:45:47,307 he has the view that is consistent with what he said he is 602 00:45:47,307 --> 00:46:10,583 interested in, which means have a lower Fed balance sheet. We 603 00:46:10,583 --> 00:46:35,975 could also have a regime change where he's going to sort of 604 00:46:35,975 --> 00:46:57,135 follow more of what the Treasury Secretary would like and 605 00:46:57,135 --> 00:47:16,179 allow for fiscal dominance. That's a different regime change. 606 00:44:52,291 --> 00:45:20,277 Andthe other third will sort of say that if we just continue to 607 00:45:20,277 --> 00:45:44,265 follow the path of these large deficits in the US, and probably 608 00:45:44,265 --> 00:46:06,254 the most interesting piece of research the last couple of years 609 00:46:06,254 --> 00:46:32,241 in the macro side, it hasn't got as much attention, is from John 610 00:46:32,241 --> 00:46:54,230 Cochrane from the University of Chicago, now at the Hoover Institute. 611 00:46:54,230 --> 00:47:16,219 So, he's written about the fiscal theory of the price level. 612 00:44:52,291 --> 00:46:01,773 Andso, he said, well, if we really want to look at this shock 613 00:46:01,773 --> 00:46:51,403 to inflation post pandemic, it's a fiscal theory of inflation. 614 00:46:51,403 --> 00:47:16,218 It's not a monetary theory. Now,the two of them were working 615 00:45:09,077 --> 00:45:37,853 together, but his view is that you're going to get inflation because 616 00:45:37,853 --> 00:46:09,027 people are going to start to buy real assets as opposed to, you 617 00:46:09,027 --> 00:46:33,007 know, debt assets if there's the expectation of permanent large 618 00:46:33,007 --> 00:46:56,987 deficits or, we'll call it, negative surpluses on the fiscal 619 00:46:56,987 --> 00:47:16,171 side. So that's going to lead to inflation. Sowhat you’re going 620 00:44:58,687 --> 00:45:19,474 to do is that you're going to put your money into real assets, 621 00:45:19,474 --> 00:45:41,860 which we've seen in the gold. If, say, you're going to put it in 622 00:45:41,860 --> 00:45:57,850 other financial assets, we're calling this sometimes a bubble because 623 00:45:57,850 --> 00:46:18,637 if too much of it goes into one type of assets. But generally, 624 00:46:18,637 --> 00:46:37,825 financial assets are being bid up because people sort of feel as 625 00:46:37,825 --> 00:46:57,013 though they may have excess savings. They may not want to buy 626 00:46:57,013 --> 00:47:14,602 real assets or they purchase goods so they're bidding up financial 627 00:47:14,602 --> 00:47:16,201 assets. 628 00:47:18,252 --> 00:47:25,376 Yeah, I mean, it's interesting because if you go back to last summer, 629 00:47:25,376 --> 00:47:32,500 I think it was Kevin Warsh gave an interview on CNBC. He talked 630 00:47:32,500 --> 00:47:40,172 about all of this. He kind of talked about the Fed being slow to 631 00:47:40,172 --> 00:47:47,296 raise rates in 2001, 2022 has been one of the, I think, biggest 632 00:47:47,296 --> 00:47:53,872 macro forecasting mistake in 40 years. And he talks about a need 633 00:47:53,872 --> 00:48:01,544 for regime change at the Fed. And he also blamed, he saw the Fed 634 00:48:01,544 --> 00:48:07,572 as being complicit in the, the deficit expansions that we've seen 635 00:48:07,572 --> 00:48:10,312 because they've effectively financed it. Andhe had a definition 636 00:47:26,147 --> 00:47:40,358 for inflation of something like, inflation happens when the 637 00:47:40,358 --> 00:47:57,727 government spends too much and lives too well, or something like 638 00:47:57,727 --> 00:48:10,359 that, which is consistent with the Cochrane perspective. 639 00:48:10,452 --> 00:48:23,465 This is consistent with this. So let me put this way. It's possible, 640 00:48:23,465 --> 00:48:36,478 if you get the Chairman Warsh, of his comments over the last 10 641 00:48:36,478 --> 00:48:47,489 years, you know, we'll sort of say the current administration may 642 00:48:47,489 --> 00:48:58,500 have no idea what they're really getting. If he's the person 643 00:48:58,500 --> 00:49:10,512 that they choose and he follows through on what he says he's 644 00:49:10,512 --> 00:49:23,525 going to do, we're going to be in a very different monetary regime 645 00:49:23,525 --> 00:49:33,535 because we're going to be cutting back the balance sheet. Youcould 646 00:48:11,770 --> 00:48:20,337 sort of say there's going to be a different view on how we're 647 00:48:20,337 --> 00:48:28,904 going to look at interest rates, and that that may not be inconsistent 648 00:48:28,904 --> 00:48:35,494 with what the Treasury Secretary would like. So, when we 649 00:48:35,494 --> 00:48:42,743 talk about uncertainty and then we talk about, okay, why might 650 00:48:42,743 --> 00:48:50,651 you want to be a trend follower? Why do you follow certain 651 00:48:50,651 --> 00:48:56,582 strategies? I don't think people fully appreciate the amount 652 00:48:56,582 --> 00:49:03,172 of uncertainty they could have. It's almost as though that, 653 00:49:03,172 --> 00:49:10,421 at different times, you see people in government and you say, 654 00:49:10,421 --> 00:49:18,988 I'm surprised by what he did. If you just read his speeches and 655 00:49:18,988 --> 00:49:27,555 read what he said, you might have a pretty good idea and you'd 656 00:49:27,555 --> 00:49:33,486 say like this is not what you were expecting. 657 00:49:35,162 --> 00:49:43,202 Well, there are different views on this. There is the, as you 658 00:49:43,202 --> 00:49:50,572 say, that's a very literal reading of what he's saying. But 659 00:49:50,572 --> 00:49:57,942 he has also said that he believes in the disinflationary effects 660 00:49:57,942 --> 00:50:05,982 of AI, and he sees scope for lower interest rates. So, presumably, 661 00:50:05,982 --> 00:50:13,352 he was emphasizing those comments when he was doing his interviews 662 00:50:13,352 --> 00:50:16,032 at the White House. Ifyou’ve seen those charts that show whether 663 00:49:43,346 --> 00:49:54,599 he was dovish or hawkish, depending on whether it was Republicans 664 00:49:54,599 --> 00:50:07,898 or Democrats in power, there is a suspicion that he might be a 665 00:50:07,898 --> 00:50:16,082 bit more political than maybe other central bankers. 666 00:50:16,172 --> 00:50:20,968 We'll say, Wall Street, and many voters, actually sort of project 667 00:50:20,968 --> 00:50:27,072 what they would like to see on a candidate or on someone as opposed 668 00:50:27,072 --> 00:50:31,868 to what they see as reality. And sometimes they actually then 669 00:50:31,868 --> 00:50:36,228 respond to what they think their constituents want. So, we'll 670 00:50:36,228 --> 00:50:41,024 just sort of say there is uncertainty here because we don't 671 00:50:41,024 --> 00:50:44,076 know what Warsh we're going to get. 672 00:50:44,592 --> 00:50:45,072 Yes. 673 00:50:45,312 --> 00:50:49,646 I think that you often see this with Supreme Court justices, 674 00:50:49,646 --> 00:50:53,980 not the current court, but if you look historically, there are 675 00:50:53,980 --> 00:50:58,314 a number of Supreme Court justices that are picked, and it 676 00:50:58,314 --> 00:51:03,436 is assumed that he was going to behave a certain way, and then 677 00:51:03,436 --> 00:51:07,376 after becoming the justice, his behavior seems to be significantly 678 00:51:07,376 --> 00:51:11,316 different than what they thought that they were getting. So, 679 00:51:11,316 --> 00:51:15,650 this is the kind of uncertainty we have. And so, what 680 00:51:15,650 --> 00:51:19,984 we're saying is be prepared for something that may be different 681 00:51:19,984 --> 00:51:23,136 than what you're reading about in the newspapers. 682 00:51:23,562 --> 00:51:27,978 Yeah, I mean, that's fair enough. I mean, maybe moving it to 683 00:51:27,978 --> 00:51:32,026 more of a quant model trading perspective. Obviously, this is more 684 00:51:32,026 --> 00:51:36,442 qualitative, but obviously, as you say, when you do get a regime 685 00:51:36,442 --> 00:51:40,490 shift, you could get shifts in relationships that would be relevant. 686 00:51:40,490 --> 00:51:45,642 So, how do you think about… And obviously the Fed is at the heart 687 00:51:45,642 --> 00:51:50,426 of the financial system. So, if there was a change in how the 688 00:51:50,426 --> 00:51:54,842 Fed is operating, whether it's a shift in towards less balance sheet, 689 00:51:54,842 --> 00:51:58,890 more active on the interest rates, or if it's fiscal dominance, 690 00:51:58,890 --> 00:52:02,202 whatever it is, that could have pretty widespread implications. 691 00:52:02,202 --> 00:52:06,618 So how do you think about that when you're maybe building models? 692 00:52:07,052 --> 00:52:32,610 Right. Well, one is that we've talked about that you want to try 693 00:52:32,610 --> 00:52:58,168 to pick up regime changes. So, we know we have the technology to 694 00:52:58,168 --> 00:53:23,726 do that. So, if we see that using, let's say, the hidden Markov 695 00:53:23,726 --> 00:53:49,284 processes, we can be able to say that if asset prices start to 696 00:53:49,284 --> 00:54:12,876 delink or have a change in behavior like a volatility, we can 697 00:54:12,876 --> 00:54:32,536 pick that up. And so, we can adjust for that. Wealso know that 698 00:52:15,024 --> 00:52:36,947 there are breakpoints in data sets. So, we have technology to, 699 00:52:36,947 --> 00:53:00,863 what we call, change point detection, to sort of see if there's 700 00:53:00,863 --> 00:53:28,765 a change or break in a time series. So, now what we could have 701 00:53:28,765 --> 00:53:54,674 is, of course, you're never going to pick a peak or a trough 702 00:53:54,674 --> 00:54:18,590 in a break point because it has to break before you actually 703 00:54:18,590 --> 00:54:32,541 then pick it up in the data. But that being said, if we're sort 704 00:52:37,005 --> 00:53:19,795 of aware that there's these structural changes, then we could 705 00:53:19,795 --> 00:54:19,701 be more sensitive to when we look at the data to see if there's 706 00:54:19,701 --> 00:54:32,538 a break point. So,what does that mean? One of the things that 707 00:52:20,902 --> 00:52:38,907 I work with my friends, with a firm we're involved with is that 708 00:52:38,907 --> 00:52:51,372 we're looking at the connections across markets. We think 709 00:52:51,372 --> 00:53:09,377 of markets as a connected system. We think of it as a network. 710 00:53:09,377 --> 00:53:25,997 And so, what we're trying to look for is the causal relationships 711 00:53:25,997 --> 00:53:42,617 between markets, And then sort of say, like, how have network connections 712 00:53:42,617 --> 00:53:59,237 changed through time? And if we see that there are changes in 713 00:53:59,237 --> 00:54:14,472 network connections, and we are willing to accept or adapt to 714 00:54:14,472 --> 00:54:31,092 these changes, then it's more likely that we could get ahead of 715 00:54:31,092 --> 00:54:32,477 it. Now,any type of trend following any type of quant model 716 00:52:24,828 --> 00:52:45,836 is always sort of looking into the past. Okay, so the question is 717 00:52:45,836 --> 00:53:10,076 how fast can we react to past data or how can we sort of react 718 00:53:10,076 --> 00:53:29,468 to these changes as they occur? One of the advantages of having 719 00:53:29,468 --> 00:53:52,092 a macro view is that when you start to see changes, or we expect 720 00:53:52,092 --> 00:54:11,484 changes, we could then be more sensitive to say, is there something 721 00:54:11,484 --> 00:54:32,492 going on in the data that we should be aware of and highlight? 722 00:54:33,432 --> 00:54:38,502 Yeah. So, is it a case that you would see a shift that is consistent 723 00:54:38,502 --> 00:54:41,882 with what you're expecting from a macro perspective? Is that 724 00:54:41,882 --> 00:54:42,220 it? 725 00:54:42,792 --> 00:55:07,380 One would hope that it's consistent with what we see in a 726 00:55:07,380 --> 00:55:29,919 macro perspective. But what we're finding, and this is one of 727 00:55:29,919 --> 00:55:52,458 the key issues that we're seeing in economics in general and 728 00:55:52,458 --> 00:56:14,997 especially in finance, is the issue of trying to identify causality. 729 00:56:14,997 --> 00:56:35,487 So, this is the number one issue in finance now. Andwe'll sort 730 00:54:47,238 --> 00:55:03,540 of say that ADIA, their labs, had their causal discovery challenge 731 00:55:03,540 --> 00:55:19,842 where they sort of said you could use different machine learning 732 00:55:19,842 --> 00:55:34,662 to look at sort of finding causal relationships. The reason 733 00:55:34,662 --> 00:55:53,928 why this is so important is that, over the last couple of years, 734 00:55:53,928 --> 00:56:14,676 what we have is what we call the factor zoom. You look at enough 735 00:56:14,676 --> 00:56:30,978 data, everybody's finding all these new risk premiums or risk factors 736 00:56:30,978 --> 00:56:35,424 in the data. Andthen what you find out is that you find them in 737 00:55:02,388 --> 00:55:23,617 the training set, or you find them in past history, then we look 738 00:55:23,617 --> 00:55:46,479 at them in the future and they no longer exist. So, there were sort 739 00:55:46,479 --> 00:56:04,442 of spurious causality, or we are trying to find something that 740 00:56:04,442 --> 00:56:20,772 we overfitted, or we found something that didn't really exist 741 00:56:20,772 --> 00:56:35,469 or it only existed temporarily in the disadvantage peers. So,I 742 00:54:45,188 --> 00:54:59,564 said, well, look, if we're having this, we'll call it the factor 743 00:54:59,564 --> 00:55:15,138 zoo, we're finding also that, well, what we find in the data sometimes 744 00:55:15,138 --> 00:55:29,514 doesn't persist through time. Well, maybe we got to go back to 745 00:55:29,514 --> 00:55:43,890 square one and say, let's look for what are the important causal 746 00:55:43,890 --> 00:55:58,266 agents. Let's look for causal factors and then see if our models 747 00:55:58,266 --> 00:56:12,642 are consistent with the actual data, or can we find causal relationships 748 00:56:12,642 --> 00:56:27,018 that we need to rethink our models. This is going to improve 749 00:56:27,018 --> 00:56:35,404 the amount overall science for quant modeling. 750 00:56:36,692 --> 00:56:41,708 Interesting. I mean, in the current context, where is that relevant, 751 00:56:41,708 --> 00:56:46,724 do you think? Or what are the ones… causality uncertain, would 752 00:56:46,724 --> 00:56:47,636 you say? 753 00:56:47,652 --> 00:57:21,108 Well, say this is a work in progress because what we'll sort 754 00:57:21,108 --> 00:57:54,564 of say, when we've analyzed some of the interesting work in machine 755 00:57:54,564 --> 00:58:19,656 learning right now, they've done some testing, some research 756 00:58:19,656 --> 00:58:55,900 where they said like, well, is it the number of features we look 757 00:58:55,900 --> 00:59:34,932 at that gives us value added or is it the complexity of our model? 758 00:59:34,932 --> 00:59:46,084 So linear versus nonlinear? Andthe machine learning world says 759 00:56:59,958 --> 00:57:26,621 that, well, you know, we want to add more complex sort of techniques 760 00:57:26,621 --> 00:57:49,182 to try to tease out relationships in data. And what some 761 00:57:49,182 --> 00:58:15,845 of the research is now telling us is that it's not the complexity 762 00:58:15,845 --> 00:58:38,406 of the technique that matters, it's the number of features that 763 00:58:38,406 --> 00:59:03,018 we have or it's the domain knowledge we have that actually adds 764 00:59:03,018 --> 00:59:27,630 more value. So, if it's an arms race choosing more complex machine 765 00:59:27,630 --> 00:59:46,089 learning techniques, we may not get as much value. So,we talked 766 00:56:57,387 --> 00:57:33,082 about peak bubble. We'll sort of say that for different techniques 767 00:57:33,082 --> 00:58:15,267 there's a while that everyone thinks that this is going to be the 768 00:58:15,267 --> 00:58:57,452 new holy grail. And then, in reality, what we find out is that 769 00:58:57,452 --> 00:59:42,882 it's a lot tougher to come up with a new model solution than we 770 00:59:42,882 --> 00:59:46,127 thought. Anotherthing we find with machine learning is, for example, 771 00:57:06,242 --> 00:57:26,691 is that machine learning is very good on stationary data. So, 772 00:57:26,691 --> 00:57:48,999 I'll use the analogy, if we're using machine learning to look at 773 00:57:48,999 --> 00:58:13,166 visual interpretations, can we find a picture of a dog? So, we show 774 00:58:13,166 --> 00:58:35,474 a hundred thousand pictures of dogs is, after a while, it learns 775 00:58:35,474 --> 00:59:05,218 how to find a picture of a dog in a photo. Okay, well, if the dog 776 00:59:05,218 --> 00:59:21,949 is changing through time, its characteristics are changing, like 777 00:59:21,949 --> 00:59:46,116 a market, that we find that it's a lot harder to do this. Andso, 778 00:56:53,914 --> 00:57:31,486 a lot of the techniques that have, in machine learning, been very 779 00:57:31,486 --> 00:58:02,796 promising in different areas of science, which has very stationary 780 00:58:02,796 --> 00:58:34,106 dictionary data, it's been very successful. When we apply these 781 00:58:34,106 --> 00:59:08,547 techniques when we're looking at time series data, that seems to 782 00:59:08,547 --> 00:59:36,726 more follow an adaptive markets hypothesis, which is what 783 00:59:36,726 --> 00:59:46,119 Andy Lo developed. Whatwe find is we find that agents behavior 784 00:57:16,857 --> 00:57:52,552 changes through time. We find machine learning is a lot harder 785 00:57:52,552 --> 00:58:21,757 to… increases predictive power. So, markets are complex, they're 786 00:58:21,757 --> 00:58:50,962 non stationary, they go through regime changes. Which means 787 00:58:50,962 --> 00:59:33,147 that the quest for finding the perfect model is ongoing and we still 788 00:59:33,147 --> 00:59:46,127 haven't solved it yet. 789 00:59:47,892 --> 00:59:51,805 Very good. So, I mean it comes back to robustness, doesn't it? I 790 00:59:51,805 --> 00:59:54,815 mean, obviously what you're suggesting is you can't overly optimize 791 00:59:54,815 --> 00:59:57,223 given the complex adaptive nature of the system. 792 00:59:58,452 --> 01:00:13,896 And when you think about, okay, for your listeners and for 793 01:00:13,896 --> 01:00:32,148 a lot of modelers, you say, well, should I tool up for machine 794 01:00:32,148 --> 01:00:47,592 learning? On the one hand, the answer is absolutely, you’ve got 795 01:00:47,592 --> 01:01:07,248 to know what's going on in AI, you have to know what's going on 796 01:01:07,248 --> 01:01:21,288 in machine learning. At the same time, your benchmark standards 797 01:01:21,288 --> 01:01:39,540 should be more simpler models. So, I don't want to sound like I'm 798 01:01:39,540 --> 01:01:50,832 talking out of both sides of my mouth. Istillbelieve in trend 799 01:00:04,427 --> 01:00:16,377 following. It still seems that it works, especially in uncertain 800 01:00:16,377 --> 01:00:28,327 regimes, especially when data is nonstationary. At the same time, 801 01:00:28,327 --> 01:00:45,057 how do I always try to try to improve on that? Because so many 802 01:00:45,057 --> 01:00:57,007 people are already trend followers, so many people are already 803 01:00:57,007 --> 01:01:08,957 using similar models. So, you're always trying to say, how 804 01:01:08,957 --> 01:01:23,297 can I look for some small improvements that are going to differentiate 805 01:01:23,297 --> 01:01:37,637 myself from everyone else and also give me a predictive edge. The 806 01:01:37,637 --> 01:01:50,782 problem comes in on margin, it's sometimes hard to do that. Now,when 807 01:00:02,614 --> 01:00:31,748 we say it's hard to do, just when I think that you've solved it, 808 01:00:31,748 --> 01:00:56,720 come up with a better model, markets may change slightly and they 809 01:00:56,720 --> 01:01:17,530 go through periods of, you know, strong performance. Then you 810 01:01:17,530 --> 01:01:38,340 get fallen performance, and then it improves. And that's even 811 01:01:38,340 --> 01:01:50,826 finding for momentum and trend following. Thereare periods that 812 01:00:04,068 --> 01:00:22,320 when it does really well, then there might be a period where it 813 01:00:22,320 --> 01:00:39,168 wanes in performance and it goes back, and this is what we've 814 01:00:39,168 --> 01:00:54,612 seen throughout history. So, right now we've had a good performance 815 01:00:54,612 --> 01:01:11,460 and trend following. Will that persist? Well, I could sort of say 816 01:01:11,460 --> 01:01:29,712 over the next five years, I could still be a believer in momentum 817 01:01:29,712 --> 01:01:45,156 and trend following. Over the next five months, that's a little 818 01:01:45,156 --> 01:01:50,772 bit harder to say. 819 01:01:52,522 --> 01:02:26,342 Who knows? Well,we'll be here, anyway, to evaluate that over 820 01:01:56,014 --> 01:02:01,058 the next five months. So, we shall see. Thanks very much for your 821 01:02:01,058 --> 01:02:05,714 thoughts, Mark. Great to catch up and get your perspective on all 822 01:02:05,714 --> 01:02:10,370 of that. I'll be back again, actually. Next week I'll be back 823 01:02:10,370 --> 01:02:15,414 from Miami in time to record with Cem. So if you've any questions, 824 01:02:15,414 --> 01:02:21,234 please get them in to us. But until then, from all of us here on 825 01:02:21,234 --> 01:02:25,502 Top Traders Unplugged, stay tuned and we'll be back again with 826 01:02:25,502 --> 01:02:26,278 more content. 827 01:02:27,142 --> 01:02:37,582 Thanks for listening to the Systematic Investor Podcast series. 828 01:02:37,582 --> 01:02:53,822 If you enjoy this series, go on over to iTunes and leave an honest 829 01:02:53,822 --> 01:03:08,902 rating and review. And be sure to listen to all the other episodes 830 01:03:08,902 --> 01:03:13,542 from Top Traders Unplugged. Ifyou have questions about systematic 831 01:02:36,424 --> 01:02:54,988 investing, send us an email with the word question in the subject 832 01:02:54,988 --> 01:03:01,176 line to info@toptradersunplugged.com and 833 01:03:01,176 --> 01:03:13,552 we'll try to get it on the show. Andremember, all the discussion 834 01:02:29,837 --> 01:02:35,766 that we have about investment performance is about the past, and 835 01:02:35,766 --> 01:02:40,617 past performance does not guarantee or even infer anything 836 01:02:40,617 --> 01:02:45,468 about future performance. Also, understand that there's a significant 837 01:02:45,468 --> 01:02:50,858 risk of financial loss with all investment strategies, and you 838 01:02:50,858 --> 01:02:56,787 need to request and understand the specific risks from the investment 839 01:02:56,787 --> 01:03:01,638 manager about their products before you make investment decisions. 840 01:03:01,638 --> 01:03:08,106 Thanks for spending some of your valuable time with us and we'll 841 01:03:08,106 --> 01:03:13,496 see you on the next episode of the Systematic Investor.