Dave Plecha is Dimensional’s Global Head of Fixed Income. Dave is an enthusiastic and incisive communicator at the firm’s conferences and seminars. A member of the Investment Committee and Investment Research Committee, he not only manages US and global portfolios but also maintains much of the fixed income research and client communications. Dave has been instrumental in the planning and execution of Dimensional’s tax-managed separate account strategies in the US.
Dave received his Chartered Financial Analyst® designation in 1996. Prior to joining Dimensional in 1989, he managed stock index futures and options for Leland O’Brien Rubinstein Associates and was an operations planner for Texas Instruments. He holds an MBA from the University of California, Los Angeles, and a BS in industrial and operations engineering from the University of Michigan.
Even among rational investors with diversified portfolios, there seems to be less known about the inner workings of the bond portion of their investments. Here on the show today to help us get a better understanding of fixed income investments is none other than Dave Plecha, Global Head of Fixed Income at Dimensional Fund Advisors. Dave is one of the authorities on the subject of bonds and is amazing at articulating the concepts at play in this arena. This conversation goes in-depth, but is also a great starting point for investors to begin thinking about this part of a portfolio, and deepen an understanding of something that is so often misunderstood or misused. As you will hear, Dave has a real passion for this subject and has been presenting and speaking on precisely this work for the last twenty years. We cover a lot of ground with Dave, talking about why his approach might be confused with a certain type of market timing, the impacts of inflation, the current low interest rates and if these affect bond investments, an explanation of forward rates, and so much more that you will not want to miss. We even find time for a quick story about Dave's early days working with Eugene Fama, so make sure to stay tuned in for that.
Key Points From This Episode:
Assessing the role of bonds in a portfolio, in relation to the current low interest rates. [0:02:30.3]
Concerns over negative interest rates for bond investors. [0:05:30.1]
Bonds and real returns; the impact of inflation on Canada's market. [0:09:30.7]
Dave's perspective on the argument for long bonds as diversifying assets. [0:15:39.4]
Comparing and contrasting the bond market with the stock market. [0:17:55.7]
The trading of bonds and what differentiates it from trading stocks. [0:22:29.5]
The volatility of last March and Dave's reflections on trading during that period. [0:27:17.8]
Differentiating Dimensional's approach to bonds from the other big firms'. [0:29:21.7]
The primary factors that influence expected returns in fixed income. [0:31:40.4]
Understanding forward rates and the information they provide about expected returns. [0:37:26.7]
Building better investing strategies using forwards rates. [0:40:43.5]
Clarifying expected premiums for maturity in a variable maturity strategy. [0:44:06.7]
Dave explains why market timing does not work with regard to fixed income. [0:46:30.4]
Quantifying the differences in expected returns from the index and Dimensional. [0:51:01.7]
A great argument from Dave for maintaining a diversified approach to all investments. [0:52:31.51]
The connection between present observable credit spreads and future realized payments. [0:55:48.5]
The game-changing development of Trace in the bond market. [0:59:40.2]
Applying the Dimensional approach as a do-it-yourself investor. [1:05:57.7]
A great story from Dave about his early days working with Eugene Fama. [1:10:25.2]
How Dave defines success in his life and work these days. [1:11:30.1]
Read The Transcript:
So our first question, with interest rates as low as they are, and this is a question we get a lot, has the role of bonds in a portfolio changed, say, to when we first met or even longer, say 20 years ago?
No, I don't really think the role has changed. I would start by saying, "Well, what is the role?" That's going to be different for different investors, but generally, the role of fixed income is going to have to do something with some sort of risk control. If we are just simply in the pursuit of returns, if we had no view towards risks, but just one on returns, well we know that in the overall scheme of things, that could easily have higher returns. So the role of fixed income is generally having to do with some risk control. So you think about some of the risks controls, asset liability matching, so you're trying to control interest rate risks.
Well, this is just as good as it's ever been, or from the standpoint of volatility, bringing down the volatility of an equity portfolio, just as good as it's ever been now. Now, yes, the expected returns are lower, but I always like to turn the focus onto the premiums, right? So when we look at any asset pricing model, we start with a risk-free rate, plus we add some premiums, presumably, for taking risks. That's true for the equity asset pricing models as well, while we're all starting with the same risk-free rate. So in some sense, we're all in the same boat, a low risk-free rate. Our premiums in fixed income, expected premiums are in line with historical premiums. But like I say, we're just in a low-interest rate world, which affects all financial assets.
How much of a premium do we expect from the term and credit, which we're going to ask you more detail about later, but if we're expecting a premium on top of the risk-free rate, what kind of numbers are we talking about in fixed income?
Well, it ebbs and flows with the shapes of the curves, and it ebbs and flows on the credit side with the amount of credit spreads. The yield curve steepened up quite a bit, at least in the U.S. and in some other curves around the world in this past year and expected term premiums went up. So say, for example, it might be on the order of 125, 150 basis points, let's say, getting out to the intermediate space in the U.S. and some other curves. Now that's up from where it was. If you look long-term historically, well, if you look historically, you get a lot of bank for the buck on the short end of the curve.
You really do. Just going from cash out to a year historically will get you maybe something on the order of 75 basis points or maybe a little bit more, and you get almost another 100 for going out to five years. So those are the kinds of numbers, and we're not too far off those numbers right now; and credit, same thing. Credit, you look at investment-grade credit in the intermediate space. Maybe historically it gets you about 80 basis points there and we might be a little bit tighter than they are right now, we're not too far off.
So when someone says that interest rates are zero, they're probably talking about the short rate, but there's still some premium out there?
Right. Premium, like I say, tacked on to a risk-free rate that is very close to zero, and in many countries, actually negative.
Yeah, I want to ask about that too. So negative interest rates, they've gotten less attention, I think, in the last couple of years than they did, maybe a few years before that. Are negative interest rates, something that bond investors need to worry about?
Well, in large parts of the world, they maybe quit worrying about them because it's been that way. We literally have on the order of $18 trillion in the world, bonds that have negative yields. So it's not like they've gone away in the Canadian market, in the U.S. market. They're not negative. That mean that they won't go negative, they may. But let's say from the standpoint of a Canadian investor, or a U.S. investor, when you look over to, let's say, Europe and you see negative yields, it doesn't necessarily say, "Oh, you shouldn't buy those because you're going to lose money on those bonds," because for one thing, if the strategy includes hedging the currency, which most of ours do, then that's going to be a great equalizer.
The currency hedge itself will equalize the short-term rates from turning the foreign short-term rate into the domestic short-term rate. So you're going to start out at the same place your domestic short-term rate, then again, we're back to these premiums again, you're going to be interested if that foreign curve is steeper than your domestic curve, that premium that you earn is going to be higher than the premium that you would earn at home, but you're tacking that premium onto your own short-term rate once you put the currency in. So negative rates doesn't necessarily mean that you shouldn't buy the bond.
Now in the end, I think, for investors that are saving, let's say, for future consumption, it's not really the nominal rate that matters anyway, is what really matters is the real rates. So I'll give you a couple scenarios. If you told me, "Well, the nominal rate is 4%, but inflation is 5%," well, if I go by the 4%, I'm losing purchasing power, if you told me, "Well, the rate is -1%, inflation is -2%," well, I'm gaining purchasing power. So it's really about, for those that are actually saving for future consumption, it's really real rates that matter more so than the nominal rates.
I'd like to ask a real basic question. Can you just explain how and why we have negative interest rates?
Well, in the end, rates come down. Like all prices, it comes down to supply and demand for, in this case, money. Now, I always describe to people, the interest rate, in some sense, is the most fundamental financial concept in the world because, let's say, "Well, what's the price of stuff? What's the price of goods and services?" Well, the price of goods and services, we measure in money. "Well, what's the price of money?" We measure that in interest rates. So it's as fundamental as it gets, but in the end, it comes down to supply and demand for anything that sets the clearing price.
Who sets interest rates?
I've talked to lots of investors all around the world, and there is so much attention people give to those central banks. I think the idea is, "Boy, if I just follow the central bank and listen to everything they say and try to read the tea leaves, then I'll know what to do with my bond portfolio." The implication is that, "Well, they're setting all the interest rates in the world." In reality, they are involved typically in a very, very, very short end of the curve; typically, the overnight rate. They're intervening where banks lend to each other, typically, in an overnight transaction. Now, there are some researchers that would argue that if you just looked at the numbers, you really can't tell if the central bank is leading the market or the market is leading the central bank. if you just look at the numbers without even the narrative behind them.
But as you get farther out the curve, it's buyers and sellers of bonds, investors, savers, like you and me, that are all feeding into this supply and demand and finding that clearing price. If you go back to Greenspan era, he talked about that too. He referred to it as a conundrum. He's doing all this activity on the overnight rates, trying to intervene in the market, he's trying to make rates go up and meanwhile, the yield and the ten-year treasury is going down. So just because short-term rates might be doing something doesn't mean a longer-term rates are doing the same thing. Curves can flatten. Curves can steepen. Curves can do everything in between.
So interesting. You mentioned the importance of considering real returns when we're talking about bonds. In Canada, we have inflation protected securities, but it's not that our market's not as developed as it is in the U.S. Should nominal bond holders in a country like Canada where we don't have a real tips market like the States. Should we be worried about inflation wiping out the already low expected returns that bonds have right now?
Well, in some sense, all nominal bond holders, when you buy a nominal bond you do bear the risk of unexpected inflation. Now, there a couple of things that I would point out on that. First of all, the nominal bond market in every market around the world, they are assessing inflation expectations every minute of every day. That's really what's going on. As economic releases get announced, people are reassessing inflation expectations, okay? So then you go buy your bond. You pay your money, your money's gone, you got to bond, you bought this fixed set of cash flows. Now, you are subject to that risk of unexpected inflation. That inflation will turn out to be higher than what it was when you expected, but that's not new to this current era. That's not new to a low interest rate environment or a high interest rate environment; it's always there.
Now, the other side of the coin is that nominal bond holders know that they bear that risk. So this is not news to them either. So nominal bond holders are going to demand a risk premium from issuers in order to be willing to bear that risk they're going to lower the price and increase the expected return by this risk premium. So if you held side-by-side a nominal bond in an inflation protected bond, the inflation protected bond does have a lower expected return by the amount of that risk premium that the nominal bond holder is demanding. Now in some markets that, investors, where they're available, some investors might say, "Look, I'm willing to forgo the premium." Now they might be gambling on rates or stuff, but even aside from making a forecast that you think the market's got it wrong, it comes down to risk aversion.
Some people might say, "Look, I'm so averse towards that type of risk, I'm willing to let go of the premium. I'm willing to accept a low return and give me a bond that's indexed to inflation." So the Canadian investor in nominal bond is no different than any other investor now, or in any other period. You are subject to the risks, but you do expect a premium for bearing that risk. The way to mitigate that, if you're more risk-averse, then within the nominal space, you could vary your duration. So the more risk-averse you are, the shorter you go, where that risk of unexpected inflation is lower because your bond is going to mature quicker and then you could get that cash invested into the new rates. If there is an inflation, those will be higher rates. Whereas the long bond, that's where you bear the most risk and that leads to a whole additional discussion about, is that enough premium? That kind of question.
So that's the trade-off, right? So you say, "Okay, as I go farther out the curve, I am bearing more risks, certainly of unexpected inflation. You see that, of course, obviously in volatility. So if we look at long-term historical returns and the volatility of those returns, obviously volatility is going to go up with longer term bonds. So you say, "I should demand a higher return for bearing that risk," and on average, you have gotten higher return. So on average long bonds have higher average returns in intermediate bonds. Intermediate bonds have higher average returns than shorter term bonds, but the rate of increase is not linear. The premium is going up, but it's increasing at a decreasing rate and that leads to all kinds of interesting questions and discussions.
One question I often get is, "What's the matter with the long bond holder? They've been at this for a couple centuries? How come you haven't figured out how to price long bond? You're systematically paying too much because the increments seem so modest." One of the things we always talk about is that question is couched within this whole mean variance framework, that while you're measuring risk by volatility, it's going up and it's going up fairly rapidly, but the premium doesn't seem to be keeping pace. That's assuming that all investors look at the market through that lens, through that mean variance lens and not all investors do. So the classic example is the asset liability mattress. So if you're running, let's say, a pension fund or an insurance company, and you have liabilities that are way out in space in long duration liabilities and, let's say, particularly those liabilities are nominal in nature so those liabilities are not indexed to inflation in any way. You're worried about changes in interest rates because you look at things in present value terms, always.
You're discounting those liabilities to present value. Well, those long bonds are a nice hedge because you line up your assets with your liabilities and then, when you take the present value of both, your assets and liabilities, now they're moving together. So in the end, that long bond to a mean variance investor is you're increasing risk as you go out. To that insurance company, he's decreasing risk. He or she is decreasing risk because they're aligning of assets and liabilities. So that might be the marginal price setter of those bonds and say, "I'm willing to pay more because it's a risk-reducing trade," if that investor had long duration liabilities and went out and bought T-bills, treasury bills, what the mean variance investor might view as the risk-free asset, that's a horrendously risky position to be in.
So there's all of these blending into the market at the same time. We're all participants in the same market and prices are getting set. So I think the starting point for investors is to understand, "Well, what is risk to me? What am I sensitive to?" Many investors would argue, I think I would, for myself, I'm sensitive to volatility, I'm concerned about that. If that's your answer, it starts to get a little bit more difficult to make the case why I need to be all the way out on the far end of the curve when the volatility is going up and the average returns are going up, but not all that much compared to the intermediate bonds.
One of the other arguments that I've heard from more of a mean variance perspective is that because the volatilities of long bonds are higher and their correlation with stocks are lower, maybe even negative in a crisis, that long bonds are better diversifying assets in a mean variance portfolio than short bonds or intermediate. Is there any merit to that argument?
The key assumption on there, of course, is that correlation. We've looked at this quite a bit. If you look at, for example, we could go back all the way to the mid-'20s` with the Ibbotson sync field data, and just take long-term treasury bonds and the stock market, in the U.S, let's say, the S&P 500. Then you can look at, let's say, rolling five-year correlations through time of those two assets. So it'd be 60 months and you roll through time and you can look at shorter and longer periods of these correlations, and the last 20 years, the correlation has been negative. But if we look over, let's say, that 90-year period, there were long stretches from, let's say, 1930 through the mid-'50s, they were positive. Then it's dip negative. Then from the mid-'60s to 2000 or a 35-year stretch, where those correlations are positive, now it's negative. So it has ebbed and flowed.
Now, in the end, if you're going to bet the whole strategy on this negative correlation, that's a big bet to make, because if you think the correlations are negative, you're attracted to volatility because you own stocks in your portfolio as well, and it starts skewing your trade off. You'd say, "Well, even if you told me the premium was zero between intermediate long, I still like the long, because I want the volatility." Even if the premium dipped slightly lower, you might say, "I'm willing to take a lower return on my long bond, because I want the volatility because the negative correlation." Looking at the last 90 years is great, but none of it matters because what matters is the next 30 or 50 years and if that turns positive, then all of those decisions you made will look very odd. So when you're forming that strategy, we know that correlations are notoriously non-stationary. They have been negative looking backwards, but we don't know what they're going to be going forward.
I have a big picture question for you. So most people are familiar with the stock market, but I think it's safe to say many people aren't nearly as aware of the bond market. How do you compare and contrast the two markets?
Well, first thing I would say is, I hear this all the time, people say, "Oh, bonds are so much more complicated than equities." I've seen just the opposite, just the opposite. For example, if I go buy a 30- year Canadian government bond and I pick a very high quality, a very, very high quality issue or their chances of default are very, very, very close to zero. Technically, let's not call it zero, but extremely close to zero. For the most part, I know my cash flows to the penny, to the day for the next 30 years. How much easier can we make it for you? If you tell me, "All right. Tell me the cash flows for Netflix or Amazon for the next 30 years." Well, geez. I have no chance, right? So for one thing, I try to tell people, in some sense, financial assets is all about taking the present value of future cash flows.
I know my cash flows in the bonds. Once they try to demystify them, the other thing is there's so much that they have in common in the end, we're used to arguments about the efficacy of markets as price setting, how people use the term efficient, how efficient out of the markets? All the studies that you're so used to on the equity side, they've been done on the fixed income side, very, very similar results, the same asset pricing notion about a risk-free rate, plus some premiums, we've got the same. Our premiums are term and credit versus the equity premium. So there's a lot that we have in common. As far as the markets themselves, fixed income has lagged in a lot of ways in transparency.
For example, we finally, in the last 20 years, got some decent post-trade transparency the equivalent of a ticker tape. So we've closed the gap there from the transparency standpoint. That eventually will lead to electronic style trading, which we're doing now in the fixed income market. So I try to point out that people that where the equity market feels comfortable and the fixed income of market feels foreign. That's a lot more comfortable than you're making it out to be. Don't be worried if someone starts throwing technical terms, convexity, and key rate durations, they're trying to confuse you. There's no need to be confused by fixed income.
How big a difference is there in size of the market trading volumes? You have a sense of that?
Well, the size of the markets are fairly similar, I think, in total, except for one. You look at governments around the world, they generally issue bonds, but not equities, so they're a huge participant now that's adding to the amount outstanding. If you look at the amount of daily trading volume, it's generally bigger. The treasury market alone is bigger than the entire U.S. Stock Market in daily dollar volume. So there's no lack of ability for price discovery, and this notion that when you have the amount of volume that's trading by participants all around the world, and then, people will claim, "Oh, prices are way off," or, "Rates are way too low, or rates are way too high and prices are distorted," it's just hard to imagine in a world as liquid and as big and now, as transparent as the bond markets are.
But there are many more specific issues of fixed income in the marketplace, like a single company might have a number of bonds in the marketplace, correct?
That's true. For any one issuer, General Electric or TD or something like that, they'll have one common equity and they might have dozens and dozens or more bonds. See, people will look at the bond market, even when you look at volume, you might say, "Oh, well, this bond didn't trade today, it's it must be a liquid," but that doesn't mean it couldn't trade today. So one difference is it's virtually all over-the-counter market in fixed income and there's not an automated quotation system, typically, like NASDAQ, where there are executable bids and offers at any point in time. But it's more of a market where if you need to sell a bond, you go out and you request a quote. You go to multiple dealers, you request a bid on the bond market forms.
If it's a liquid bond, you get lots of bids. They're all very tight with each other. You do the trade and then the market cap goes away. So it's because you don't see volume doesn't necessarily mean it's not liquid, but that's what happens when it's fragmented, like you said, over thousands and thousands of different securities, many times dozens, if not hundreds, from one issue or at all different parts of the curve and all different currencies around the world, it does get more fragmented, but that doesn't necessarily mean it's not liquid.
Can you just talk a little bit more about trading, because I think in a lot of cases, people who are investing their own money, they have maybe bought a stock or maybe they've bought an ETF, but I would say most people listening to this podcast, for example, have probably not traded a bond. Can you talk a little bit about the differences?
One thing I would say for most individuals, if you were to go buy 100 shares of XYZ Corp, the equity, you're going to buy on the offered side. In the modern area, now that they're doing even fractional shares, this notion of really, really getting penalized for trading in an odd lot on the equity side is really much, much diminished compared to where it might've been 25 years ago or more. On the fixed income side, if you're trading in very small volumes, in the institutional world, anything below a million bucks, we would call an odd lot. But if you get below, let's say, $100,000, all of the research done on transactions costs in fixed income, they're super sophisticated, but they point to one dominant factor that says that retail investor is doing some odd lot trade 10 grand or 20 grand is getting killed.
They are paying way too much. If they're selling, they're not getting near enough. On the order of 100 basis points compared to where someone who's at institutional investor, we would argue, you need to have a wide network of way seeking liquidity, liquidity to buy, or liquidity to sell an extensive dealer network. I think you would want to be tapped into primary markets in addition to secondary markets. Now in the last several years, you want, I think you want to be tapped into the peer-to-peer networks where you're not even going through a dealer, you're in these systems where you're just trading directly with each other on an anonymous basis. You need to have all of those channels developed to give yourself the best chance to get the best execution. If your banker calls you up and says, "Hey, I got a bond. I think you ought to buy it." "Well, where are you offering it?" "Well, I'm offering it here." You say, "Okay, well, that sounds good." You're probably costing yourself quite a bit in doing the trade that way.
That's really interesting. I remember reading a few years ago about liquidity in the bond market, apparently drying up because of some regulatory changes that had happened. Can you talk about what those changes were, and if that liquidity is still a concern or if it was ever a concern in the bond market?
Well, the regulation, I think, that you're referring to is really not in the investor side. It's not the mutual fund side or the ETF side, that side. It's mostly on the dealers where things have changed a little bit. You go back to 2008, a lot of these dealers, which are big commercial or investment banks, a lot of them got bailed out by governments and regulators came in n, "So now that we've saved you, there's going to be some new rules for you here." You know what I'm saying?" Generally, the rule was you required more capital, if you hold risky positions and your inventory, we're going to make you have more low risk capital to provide a buffer against the potential losses, and tying up capital is a cost. So there's cost because carrying inventory became much, much more costly. So dealers, if you look at the total dealer inventory of bonds and like in the U.S., they have to report their aggregate positions into the New York Fed, so we could see the data.
Well, the total size of the inventory came way, way down, literally by 90%, over the ensuing, let's say, decade. So they're holding much, much smaller inventory, but then you look at trading volume. People say, "Well, that must have dried up liquidity." We could see trading volume in trades. Trading volume has gone straight up. Every year, as it's higher than the year before. By the way, people talked about March of 2020 as if there's some liquidity crisis.` You look at the amount of volume, all-time high for the amount traded was in March of 2020.So let's put those two facts together. We have dealer inventory way down, but volume way up. So if you think of that normal channel seller to a dealer, dealer to a buyer, and you measure something like what we would call a inventory velocity, how many tens are turning over there and know the velocity is way, way up. Dealers are turning over their inventory a lot faster, holding onto it for a shorter amount of time.
Their role has changed. Often, all they're doing is they're playing matchmaker. You're a seller of a bond. They want to take it to the inventory, but they'll go find you a buyer and then they'll connect you and then the trade will go through them. But if that's what their role has become, that's what opens the door for this peer-to-peer world. If you say, "Well, all you're really doing is just, you're not positioning it. You're not owning it your own inventory, you're just finding a new buyer." Say, if I find my own buyer, the networks developed. Again, I was totally anonymous. I don't know who that buyer is. They don't know who I am, but there are ways now, who needs the dealer. Let's just find each other directly and whatever they were making in their markup, we cut that out and we could both be made better off.
You mentioned last March, which was a very interesting period for fixed income trading. Can you talk about what happened when you said the volume was very high, but there was a lot of very volatile movement around a lot of ETFs? How accurate do you think those pricings were in the ETFs, and how do you view that trading period?
Well, first of all, volumes were high, but volatility was extremely high, right? So the world is coming to terms in a very short amount of time with this pandemic and all the ramifications of that. Major chunks of the global economy are going to be shut down, and we're all figuring that out in a very narrow window of time. So volatility is spiking. Bid-ask spreads are widening out. We're all figuring that out, so you have all of this. So just because volume was high, doesn't mean that volatility can't be high and bid-ask spreads can't be wide, because all of that was happening at the same time. Now, then we saw ETS, for example. They have a NAV and then they have a price of their shares. We start seeing dislocations and people start making the claim, "Oh, the NAVs are wrong," because the true price discovery is happening in the ETF share and that's not clear that that's the case.
We look at our entities and the prices that go into the entities and these and we're pretty comfortable at Dimensional with all the entities that we publish. We're quite comfortable with those. The way I might frame it is if you go to the market, let's say, you own a bond and you say, "Well, if I go to the market to sell a million of it, what's the price?" "Well, you're going to get a price." "If I go to the market to sell 50 million of it, what's the price?" "Well, it's going to be a different price." If you demand immediate execution on both, 1 million or 50 million, you're going to get different prices. The ETF market people are going, "Well, those $50 million orders, they're the big orders, and what you were seeing was market impact, which is different than price." So, because we saw differences in the NAV and the ETF share price, doesn't imply that either are necessarily wrong. One is reflecting essentially a market impact that the other one wasnt.
How does Dimensional's approach to trading bonds compare it to other big firms like BlackRock or Vanguard?
I can't speak specifically to those, but I could give you some general. So, at Dimensional, what we really, really emphasize a flexible approach and portfolio formation, a flexible approach at the time of trade. We're trying to deliver these premiums, these terms and credit premiums. For every bond we buy, there are lots and lots of candidates that will have similar characteristics that are going to help in the same way to deliver those term of credit premiums. So we're not married to one bond when it comes time to go buy. So some shops, for example, I'll give you two ends of the spectrum, the active shop that has a very concentrated positions versus the index. Well, the active shop, they think that they found the best pricing and the issuer may be in industry they've got a much more concentrated point of view on what they want to go trade.
For us versus them, they decide, "I want to own this issue," or they're going to go buy that issue of if we go to buy that issuer and the offering that we're seeing, we don't like the levels, we say, "Fine, we'll just cancel that, or I'll give you a different one different issuer, but similar characteristics," that puts us at a nice bargaining position. I always tell people that the most powerful position you can put your trader in when you send them out to the market with an order is give them the ability to walk away from any trade. It's like when you're buying a house, once you fall in love with the house you're done, as far as you're negotiating. On the other end of the spectrum is the indexer.
Now they share some characteristics of a highly diversified portfolio, but they have very, very specific characteristics that they're going after because their objective in life is to minimize tracking error. They might not be as flexible at the time of trade, so we are very intense about our credit transactions cost analysis and what we do is we look at where we traded a bond versus other people trading that same bond on the same side of the market and roughly around the same timeframe. We see ourselves being able to transact, both buy and sell, somewhere in the order of, let's say, 10 to 15 basis points better than our peers in the market, in the same security at the same time. I think it's all because of that flexibility that we give ourselves and how we form that portfolio and how we do the trade.
I come back to factors. You mentioned term and credit briefly earlier in our conversation. Can we just dive into what are the factors that explain differences in expected returns and fixed income? We also understand that you've done some recent research on factors and what works and what doesn't, if we can talk about all that, that would be great.
The two that we've identified as primary drivers of fixed income returns, term and credit, those are not controversial. They're surely not meant to be controversial. It's pretty straight forward. So you pick out a single issue where, let's say, the Canadian government and you look at Canadian government bonds and you form portfolios of short-term bonds or intermediate term bonds, long-term bonds and you let them run through time and you see differences in returns. It's the same issuer across the whole spectrum. Obviously, it's the difference in term that's driving the difference in returns, or you can go the other axes, you can say, "Well, I'll hold a bunch of five-year bonds. They're all in the exact same term. Some of their government bonds and some are single A bonds and are BBB bonds, and you can see differences in return along that axis as well."
So term and credit are the main drivers of returns. Our research all shows that in the end, the current price of the security is reflecting all the available information. There's information about those expected premiums that we could get from the current price. We'll get into this, I think, a bit later on the discussion, I suspect, about forward rates, but I'll just use the term now that ultimately taking the current price, we will calculate the forward rate, which becomes our proxy for expected return. We've been doing that for decades now, going on, well, over 38 years. Recently in the last couple of years, we've seen a lot of other factor type strategies, out in practice, a lot in the academic literature, people are talking about other factors. So naturally we're interested. We want to test those factors, but we realized the right way to test those is that we say, well, let's say you you have a factor and you might say, it's value you some measure of value, just like they did on the equity side.
That was common in the fixing of more people, saw what was going on the equity side and they said, "Oh, let's apply all the same factors on, on the bond side." So you got to figure out how to define value and then you look at the returns and you say, "Wow, these value ones seem to have higher average returns." Well, we realized that, "Look, at the start of the period, we already had an expected return, that was that forward rate. I had that at the start of the period. So if you show me the end-of-period results, what I want to do is I want to subtract out from the end-of-period results, my start-of-period expected return." Then, what I'm looking at is the unexpected return, this residual. Now, if I find something systematic in that, that tells you, "Well, your starting period expected return could be improved on because you've missed something that was captured in this other factor."
Well, when we went through all these other factors and subtracted out our start-of-period forward rate, our expected return, there's very, very little left that's systematic. In all of these, I think we studied a total of 14 factors, we did see something, something in very short term equity returns. We looked at equity momentum. We looked at bond momentum. We looked at credit momentum, a lot of these factors. What we did see is that for companies whose very, very recent equity performance, let's say, was very poor, let's say, in the bottom quintile, their corporate bonds tended to slightly underperform the following month. Now, it's very short term in nature, so it's not something that we said, "Oh, we should sell it, all those bonds, and then a couple months later, buy them all back because the effect is done or try to chase it in the other direction." But we can do as much like how we deal with momentum on the equity side.
We say, "Well, what we can do is avoid trading." If a company stock really had a bad run in the last month, we could say, "Let's just hold off on buying those corporate until that smooths out, or if it's going in the other direction that's had a really positive run, if we were otherwise going to sell the corporate bond for other reasons, let's hold off a little bit and sell it a bit later." So we did learn something, but I think the important thing we learned was how good forward rates through turns out to be. So it's, we don't rest on our laurels. We've been using for decades and decades, but it doesn't mean we don't challenge them. We want to keep proving to ourselves that we're on the right track with these things. It was really impressive how well those rates held up, because once you took those into account, all these other factors pretty much went to essentially zero.
You just said that stock momentum helps predict fixed income returns n can be used as part of a trading strategy. That's kind of mind blowing
Well again, well, when it's very, very short term, I wouldn't use the word momentum because momentum implies multiple periods and we're talking about just a single period. If you defined the month as the increment of time, which is very common and research is found in the most recent month, so we call it short term equity returns and not momentum. Even on the equity side, even momentum itself, in some sense, it's a head-scratcher trying to explain theoretically why it's there, but you can't deny that it's there. What you do is if you form your strategy where you say, "I'm not going to chase it." So you say, "What's the worst case scenario?" It all goes away one day.
We can't explain, but then one day it goes away, but I didn't do anything that would cause me harm by acknowledging that it's there. I avoided some trades, but I didn't go out and create new trades because of it. So it's definitely in the data. We're going to take steps. If it does turn out that the corporate bond will underperform. We will have held off buying it for a bit. It's a very short term thing, by the way, and we refresh that data. We refresh it every day, so we're looking at a trailing 30 days worth of data.
So interesting. I want to come back to forward rates. So I'm going to make a statement and then I'm going to ask you to explain it. In the '70s and '80s, Fama's research found that implied forward rates from the current term structure provides reliable information about expected returns across the duration spectrum. Can you explain what that means and why it's important to a fixed income strategy?
In 30 words or less explain forward rates, in other words, right? So, all right. So I've been doing this for a long time because it confuses a lot of people and so I've come up, over the years, with a little simple example. So I hope it helps. I'll lay it out there for you. I'll give you an example. Let's say you have a two-year horizon and the only bonds you can choose from are one-year bonds or two-year bonds. No, let's make it simple and we'll use zero coupon bonds, so I know everything we're serving. So I could go buy, right now, a two- year bond and I can know with certainty how much money I'm going to end up with. Okay. Now, my only other choice is a one-year bond, or I could go buy a one-year bond and I'll know with certainty how much money I'll end up with a year from now.
But then I'll have to go buy a new one-year bond to get to my two-year horizon. But here's what I know. I know how much money I could have with certainty in two years, and I know how much money I could have with certainly in one year. So I could solve for, "What does the one year rate have to be one year from now to make me indifferent between those two?" So I get, it's just a simple division there, and that's a forward rate, the one-year rate starting in one year. So that's easy math to solve for the forward rate. Now, for years and years, people would look at that forward rate and they're asking, "How should I interpret it?"
A lot of people concluded, "Oh, that's the market expectation for what the one-year rate will be one year from now." That's how the market determined what the rate of one-year and two- year bonds ought to look like relative to each other, because they're expecting the one-year rate to be this. Well, pharma was among the first to just start doing empirical tests. Are those forward rates good at predicting what the one-year rate's going to turn out to be? You just go through the data, you formed the forward rate, you move forward in the data and you look, and then you find out, "Oh well, they're terrible predictors of what the one-year rate is." So forward rates were not good predictors.
So what he found was if you want a better predictor of the one-year rate one year from now, just use today's one-year rate, just use the information price in the current curve. Now, that gives me a third alternative now. I gave you two alternatives and how to spend the next two years, but now there's another one. You can say. How about this one? I go buy a two-year bond. I hold it for a year. I sell it. I take the proceeds, go buy a new two-year bond, hold it for a year, and I sell that and now my two years is up.
With the way the math works, if you say, "Well, the expected return of that buying the two-year and holding it for a year and selling it, and I told you use the current one-year yield as your proxy for the one-year yield a year from now, it's the exact same math as the forward rate, you calculate the exact same thing." So now, saying the current one-year yield is a better proxy is another way of saying that forward rate is a better proxy for expected holding period return than it is as a proxy for the future one-year rate. So that's the crux of the final research. So now we have a way of calculating expected returns for bonds held to maturity, which is pretty straightforward. Yield to maturity is a good way to go, but now we have a way of calculating expected returns when I don't plan on holding a bond to its maturity and that's the forward rate.
Okay. So with that out of the way, how can that be used to build a smarter investing strategy than just holding, for example, cap- weighted bonds?
So you say, "Now I have an expected return model." Okay. So let's say that our two-year example, again, "Well, which one should I do?" Well, the answer is, "Well, it depends on what the curve looks like." Now, by the way, in that third strategy that I laid out, two times I had to go sell a bond. So I do have to take transactions costs new account. The other two strategies I was able to hold to maturity. But if the curve is steep enough and the transactions costs are low enough, then if the curve is upwardly sloped, typically, while the highest expected return is going to be that buy-to-sell at one strategy and just do that twice. If the curve was perfectly flat in the absence of transactions costs, all three would have the exact same expect return. With some transactions costs that that third one would be slightly lower. If the curve is downwardly sloped, well, the highest expected return is going to be by that one year end hold to maturity. Just do that twice.
So now, not only do I have expected returns, I understand that expected returns, they're not constant through time. They change their time and where I want to position the portfolio is not going to stay the same. The day I buy a bond, I might not plan on holding it to its maturity. I might plan on holding it through a segment of the curve, or as the curve changes, once I consider transactions costs, I might be willing to move from the bonds I had to some different bond that has a higher expected return. So it allows this systematic approach of going after and positioning the portfolio for a higher expected returns, but not because you're forecasting some kind of change in the curve, it's you're just using the information in the current curve, in the current prices.
So practically speaking, this would be you're holding bonds of different maturities based on the shape of the yield curve at that point?
That's correct. Now again, so I gave a simple example, now let's start spreading this out. We're not just looking at two-year horizons or bonds, we're going across the whole curve. We're adding in all the corporate bonds. We are looking at a dozen different curves. Next thing you know, I got 18,000 bonds I'm doing this on. Then, we're going to calculate expected returns for a whole array of different holding periods. Literally, we're doing hundreds of thousands of expected return calculations every day and then what it really comes down to is it really comes down to that expected return. What do you think about that expect return, that forward rate, because we're going to always be positioning the portfolio towards that higher expected return based on current prices. But I always tell people, "Look, nobody's more interested in that than we are. We're testing that and challenging that nonstop, and we've been doing it for 40 years."
It holds up pretty well. In the end, we know that there's going to be noise when I say the words 'expected return,' think of that as the mean of a distribution and there's, and it's pretty symmetric, right? So there's going to be a left-hand side and right-hand side, negative noise and positive noise. We know that there's going to be noise. Some of the periods, it'll be negative. The actual return will be short coming, less than the expected return, and some, it will be positive and it will get more than expected. But in the long run, if your expected return model is good, all that noise cancels out and it ends up as a big fat zero when you average it out and positioning the portfolio for higher expected returns will lead to higher realized returns.
You mentioned at the very beginning of our conversation, the premium for maturity, when we're talking about a variable maturity strategy, is that expected premium the same as saying what the maturity premium is, or is it different? Higher?
Well, again, the expected premium is going to be a function of the steepness of the curve. So when we talk about premiums, you're saying the amount of excess, let's say, the risk-free rate, how much am I getting over the T-bill? And flat yield curves, that's the thing that maybe is a little bit of a struggle because on the equity side, you're used to this notion that I always think small has a premium, every day. I think smaller to premium over large or value as a premium over growth. I know there's lots and lots of variants and not every day or month or a year, or maybe even decade, in some cases, that you don't collect those positive premiums. But there are times in fixed income where I see a flat curve, I don't expect any premium, or if I see an inverted curve, I think those we expect a lower return for those longer bonds than the shorter bonds.
So for us, that's really the heart of their maturity. If I'm expecting a lower return, and let's say I'm sensitive to volatility, and lots of investors are, why do I want more volatility and lower expected return? I don't. I'll move to the shorter end, at least and increase my expect return and decrease my volatility. It doesn't mean that the market doesn't know how to price assets. It doesn't mean that the world is going to spin off its axis or something like that. It just means that there are times where there's a greater demand for longer bonds and shorter bonds, so relative to supply and the curve could flatten, or even invert.
I wake up every day and expect a positive value premium, but you don't wake up every day and expect a positive maturity premium.
Not if the curves are inverted, I don't, and we take action as a result. So if you look at a more static portfolio, let's say, an index. The index, they're very, very sophisticated. It's very quantitative, but their fundamental objective, and I'm an old modeler from way back from my college days the objective function you would say, has to do with minimizing tracking error, right? So the shape of the curve doesn't really matter. What matters is, does my portfolio track? For us, the shape of the curves matters a lot because we're out trying to maximize or increase as much as possible, the expected return subject to all the constraints that define the portfolio from this objective. But the shape of the curve and expect returns matter a lot to us.
Okay, Dave, you got to help me here. So I'm sure there's lots of listeners right now that are scratching their heads, listening to this saying that this is market timing. This is active management. Ben and Cameron talk about passive ETF index type philosophies. What do you say to that?
Well, one way of answering this, and I'll give the full answer, but one way of answering it is I would argue this is the lack of all market timing, because what we're doing is we're looking at current prices and we're looking at the current yield curve, and we're calculating a set of expected returns without forecasting any changes. We're just looking at the curve. This is based, of course, on decades and decades of research. Then, we're choosing a place to be. Now, what happens if you choose anything other than that? You're choosing something other than what the current curve is offering is the highest expected return. It must be because you have a forecast, right? So a lot of people look at an upwardly sloped curve and they say, "Yeah, but I think rates are going up. So there's no way I'm going to take a net duration out the curve.
I'm going to stay short because when I'm right in my forecast, rates go up, I'll be so glad that I had short duration relative to my benchmark, and I'm going to outperform my benchmark. We're not going to do that. We're going to say, "The current curve is offering this and we're going to go with the best alternative from the current curve. Now, I think what throws people is, well tomorrow, the curve might be a little different, and the day after that, it might be a little bit different. Then they see this activity that, "Oh, they were here, but now they did a trade, and now they're somewhere else now." For us, we're saying, "Look, we are taking to account transactions because we're very intense about studying that. But given the new curve net of the transaction costs, we think we're going to be better off at point B instead of a point A." So I wouldn't confuse the activity or the turnover with market timing. I would say it's a disciplined approach to taking what the current market is offering.
So let me ask you a similar question, which we often get, which I find amazing people will come and say, "Okay guys, I get it. Mark is efficient. I want tilted portfolio. I want the premium, small cap, whatever value I get all that the market works, but on fixed income, let's face it. Interest rates have to go up. Inflation has got to be calming. Therefore, what are you going to do? You got to be effectively the same. You have to be active there. So does market timing work in fixed income?
There's absolutely no evidence that market timing works. Anyone who's ever refinanced a mortgage thinks they know how to forecast even though they'll be absolutely would never try to time the equity markets. When people tell me their story, "Oh, I think the fed or the central bank is going to do this, and commodity prices are doing that," my first question is, "Well, what part of that story do you think is not already priced in?" If you and I could talk about it, don't you think every other investor on the planet knows that too, but somehow you think there's some delay that it hasn't worked its way into prices yet, if is that common knowledge, if it's a foregone conclusion that yields are going to go up, they'd already be up. So why aren't they up? Anyone could sell their bonds anytime they want, but yet, this is where they are.
It's so interesting to think about that. Doing something different from the index is not necessarily active management when it comes to fixed income.
I agree. Like I say, it comes down to your objective function. With our expected return model, by the way, not only can we calculate the respect return of every bond in our universe, well, I could also calculate the expected return of every bond in the benchmark, in the index. I know the weights, so to get the portfolio or index expected return, you just multiply weight by expected return and you add them all up. So I know the expected return of our portfolio. I know the expected return of the benchmark. We are forming portfolios with higher expected returns. So we're not trying to tie the benchmark. We're trying to beat it, but we're doing it in a systematic way that we're not starting out by saying, "There are prices that are wrong, or the market's got the levels and rates all wrong." It's just taking the information that the market is currently pricing in and saying, "I can form a portfolio of the higher expected return rather than the benchmark." I'm out to beat that benchmark.
There's this little nuance between, we're saying you can make changes to your maturities based on the current yield curve, and that's not market timing, but someone who tries to make changes to the duration of their portfolio in anticipation of a yield curve change, that is market timing.
I think that's right, because just go to the fundamental philosophy behind it, in anticipation of the change is basically saying, "You know what? I don't agree with current prices because I think they're going here and I got to get there before the rest of the market catches up with my thinking on this." Our current thinking is, "No. I think prices are pretty good. So now here's what they are. What do I want to do about it?"
Are you able to quantify, like you said, the expected return of the index and the expected return of a portfolio that Dimensional is managing, are you able to quantify what those differences in expected returns are?
Absolutely. We are. We do that now. I will say this, that when it comes to like putting that on our website and that kind of thing, our compliance group, and I think for very good reason, they worry that when people hear the words expected return, now I'm speaking of that from a completely statistical point of view. I'm speaking of that as a mean of a distribution and the standard deviation of that distribution as a function of the duration and the credit line and all kinds of things. I basically know that at the end of the period, if the actual return turns out to be the same as the expected, that would be a miracle because they're seeing it. So I think they really worry, and for good reason, about the sharing, "Oh, this is the expected return," because all you're going to lead to is a bunch of disappointed clients, "Well, you told me to expect this, but then I got that."
So when I keep using that term, and it's a very statistical term that I'm using, which means, which really, in the long run, in the long run, and I'm not depending on the duration of the portfolio, the longer the duration, the longer the run I'm talking about, you will see the realized returns converge to the expected return over time. But in the short run, it's highly unlikely that the actual return will turn out to have been the start of period expected return. That noise, whether I end up on the right hand side of the mean, and the left hand side, that I don't know. That's something I just don't know before we start.
Absolutely. We talked earlier about the size of the bond market. When it comes to stocks, I think it's becoming at least more generally accepted that you need global diversification if you want to capture expected stock returns. So I would argue that it's pretty important. In fixed income. It's much more common to not have global diversification, at least from what we see. How important do you think it is to be globally diversified in your fixed income allocation?
I think it's important to be diversified in all of your investments. One of the things I have been thinking about is our founder, David Booth, he talks about these key cornerstones of investments, the big ones, and he said, "They only come along about once every decade. It's not like the marketing group is calling for one every quarter. That's not how it works," and certainly diversification is one of the cornerstones on investing. There's no question. That's, I would argue it's important within an asset class. It's important across asset classes. It's important around the world, globally diversified. Now what I think often throws people at first on the fixed income side is the exchange rate risk.
So if I buy outside of my home country and all of a sudden I don't hedge the currency, and I started looking at the volatility, you say, "Wow, this is way more volatile than if I just bought my home country's bonds." The whole point of diversification. Again, I'm speaking statistically now is you'd like to see lower standard deviation, non-perfectly correlated assets blending together. You don't get that if you don't have the currency. So for those that are saying, "I want a diversified portfolio," which is potentially a lower standard deviation, then you should hedge the currencies and be globally diversified.
The other thing that I would say, so that's a big risk control, risk management argument, but also, it gives you more avenues to seek these premiums. So we talked about in flat curves, we don't expect a premium. So if you only do it domestically, you give yourself one curve to look at and that's fine. So if the curve's steep, you go out and you expect a premium. If the curve is flat, you shorten up and you wait for it to get steep again. If you invest globally, you can have eleven other avenues. If any curve is upwardly sloped or steeper than yours, well the expected premium is going to be higher than yours. Let's say, they're all equally steep, well, now we get into the diversification that if they were all equally steeped and had all the same premium, I would much rather hold 12 of them than just one of them, because I know they're not perfectly correlated.
We talked about the premiums and we've talked about how to capture them. If someone just owns a universe bond fund, like a cap-weighted universe bond DTF, are they still getting access to the premiums?
They are. They're going to get the long run average premium for term and credit. What I'm saying is they could do better if, instead of taking a static approach them, because what you're going to find in those portfolios, the duration is going to be pretty stable. It's not really about anything to do with curve, shapes or expected returns, it's about composition. So it's typically going to be, "We'll hold all the bonds from one to 30 years or something like that, investment grade something like that, a global aggregate type benchmark," and so that's going to be very, very static and it's exposures determine credit. We would argue that if you take a more dynamic approach and again, we're going to take into account transactions costs. We know that it costs to get from point A to point B, that we're going to keep those as low as we can and make judicious decisions about moving and the more dynamic, you'll end up with higher average premiums than a static approach.
That just triggered a question I've been meaning to ask you for a long time. So if you look at a broad-based index fund that may own corporates, as well as government bonds and in a period of crisis where the corporate bonds may go down in value and have higher expected returns and index, correct me, but I believe the ETFs will follow the index and have a lower weighting in those corporates. You may choose to have more corporate. So does that effectively become a value play on the corporate fixed income? Is that a fair way to represent that?
I don't have a problem with it the way that you frame it. I really don't because here's what we know about credit premiums, that there is information about expected credit premiums in current prices. What we're going to do is we're going to look at the current price. We're going to convert to a yield. We're going to subtract off the yield of the treasury bond with the same duration, and we're going to see a credit spread. Then you look through time and you say, "Well, the credit spread, sometimes they're wide and sometimes they're narrow," so we see an ebb and flow. Then, the research will all suggest that there is information about realized credit premium, so the credit spread is what I observe at the start of the period. What I really care about is, "What do I get to keep at the end of the period," because there might be a bond that goes bankrupt.
"Well, it had a big wide credit spread, but I didn't get any of it because the bond defaulted on me," that kind of thing. So is there a connection between today's observable credit spreads in future realized premiums? The answer is yes. In what it is when those credit spreads are wide, on average, you do capture higher credit premiums. Well, how did the credit spread get wide? Now I'm not going to use the word value, although I understand where you're going with the question, but prices are relatively low. So that's where your word value comes in. Prices are relatively low on those corporates, which lead to wider spreads, which will lead to higher average credit premium. So our approach, we call it variable credit, will vary the amount of credit versus government based on those credit spreads.
When credit spreads are wide, we're leaning more into credit. Now everything's got limits by the way, have limits on the amount of credit. We have limits on the amount of BBB or single A or AA, everything's got some limit, but we tilt more towards the credit and tilt more towards say, the lower end of those BBBs when spreads are wide. When spreads are getting narrow, there's still a premium. I'm not saying you don't get a credit premium, but the credit premium is smaller.
So then you'll, ultimately, have to ask yourself, "Am I willing to take the same amount of credit exposure in narrow spreads or wide spreads? Am I going to be static or am I going to be dynamic?" Our view is, "Look, when the expected premium is high, we want to take more; when the expected premium is low, we're not as willing to have as much." So that variable credit approach, again, when you look at the results, it will lead to higher average realized credit premium than a static approach. You say, "You know what? I'm just going to pick a number and I'm just going to stay there through thick and thin."
It's effectively moving. I think this is what Cameron was getting at, it's moving in the opposite direction in terms of expected returns relative to the index.
Yeah, so you think about that. I know, Ben, it worked that way on term too, and I'll cover both. Why our credit spreads widening because people are still not accredited. They would rather hold the government than the credit so that relative the prices is going down and therefore, the credit spread is widening exactly when we're saying, "Let's go buy credit." So when they're selling, that's when we want to buy credit or let's think about term, why is the curve steepening? Well, people are selling out of those longer bonds because they'd rather unload the shorter bonds in the longer bonds where they had whatever fear, inflation or whatever they're selling, that's exactly when we want to say, "Oh, the curve steep. We want to take more duration now."
Well, I've heard people call it, it's a contrarian point of view. Well, you can call it that because we're often buying corporates when others are selling or buying, moving out the curve when others are shortening. But it's not because, again, we're focusing any changes. It's just, this is what the curve is offering. If the curve steepened, the longer end got cheaper relative to the short end. Now it's from an economic return standpoint, it's become more attractive to us.
I appreciate you guys improving on my question. That's awesome. Can you talk about trace? You mentioned it earlier. Can you describe what trace is and why it was a big development in the whole bond market?
It was a game-changer for us. Trace is, basically, you're so used to the ticker tape and the equity, I think about those old black and white movies and The Great Depression, they were following the ticker tape and then they look at it n then they jump off the side of the building. That's how the movies went. Well, that ticker tape has been there for 100 years. I literally Googled this and I was trying to find, when did the equity ticker tape start? They traced it back to the mid-1800 with the telegraph machine, and they talk about even before that, there were messengers on horseback. So this idea of a trade happening and that price getting disseminated to market participants, the price of the trade that goes back to the dawn, practically, of equity markets.
That concept didn't hit the bond market until 2002, which is just staggering. But that's really what trace is. It's like a ticker tape when a corporate bond trade occurs that the information, not who the participants were that, and I don't think they should, I think that should remain private, but what the bond was, the quantity, up to certain limits, but generally the quantity and the price get disseminated out to the marketplace. So it was our first real, I'll use the term post-trade transparency that we had into the market and it was a game-changer. Prior to trace, and I started in this at Dimensional in 1989, prior to trace, someone would call you up and say, "Hey, I got these five-year Coca-Cola's I could offer to you at this credit spread, 50 basis points above the five-year trade. That bond might've traded 15 minutes ago. You had no way of knowing.
If it did trade, you had no way of knowing what price it would have traded at. Then as you're on your own, "Well, it's five-year Coca-Cola's at a spread of 50, is that good? The credit spread is at 50?" Well, we had ways to survive and protect ourselves and thrive in that environment. Well, all day long I'm listening to bids and offers and I saw Shell Oil a little bit earlier and that's similar credit or someone bid me on my three-year Coca-Cola, so that anchors my thinking. So you're in the market and you're operating that way, but when there's nothing that could match seeing the trades, "Okay, Coke traded. I could see it. I can see the price. I could see the credit spread. Now I could see all that offer of 50 is terrible. That ought to be a 55 because it just traded there." It was really a game-changer for us.
So if it was a game-changer, how did it actually change the way that you approach fixed income?
Well, I wouldn't necessarily say that, in some ways it changed our approach, but mostly it changed where we can apply our approach. So in a world pre-trace, people would say, "Oh, the market was opaque. I always say, "I wished it was opaque. Man, it was black out there. I saw nothing." But what are we willing to do in that type of market? Okay, well, we're willing to do essentially high-quality A, AA bonds, government bonds, where we would look at broker prices on those. We're monitoring the credits that way, but that's where we knew we could be successful. As trace happened and there's the post-trade transparency, now we can start looking at other bonds.
Researchers could start doing their thing and start studying credit premiums, which led to this variable credit approach. But now all of a sudden, we start saying, "You know what? We're ready to go, single As, BBB. Later, we're ready to go BBs, even later, a certain mortgage type instruments, TBAs to be announced, a report on trace." You give us transparency and it'll show us where things are trading, reliable transparency that will lead to research and that will often lead us to be willing to say, "I think we could perform successful strategies in this new part when there wasn't transparency.
There are also bond rating agencies that help you understand the credit of a bond. Does trace match up with agency ratings of credit?
That's a great question. When I think about probably the three of the most important things how we use that trace data, one of them is credit monitoring. The other one, just quickly, is we model the dealer inventory. We make our own estimates of what we think is out there in street inventory. We also do this intense transactions, cost analysis, using all this trace data. But credit monitoring is as important as anything that we do. So what we're looking for, we're looking at this trace data. Now we get over 40,000 data points a day and we don't even wait to the end of the day. We're literally going in 20-minute increments grabbing the most recent 20 minutes where the price is converting price-to-yield, yield-to-credit spread, mapping those out against the yield curves, we form FRED curves. Here's the AAA credit spread. Here's the AA, a whole curve, and then we're looking for outliers.
So let's say, for example, you see this bond. Well, the rating agencies say it's single A, and everyone says it's a single A, but when I look at how it's trading and I look at it, and these are actual transactions, buyers and sellers meeting at a price, and it's mapping out right on the BBB curve. Now you're going to eventually get to a fork in the road and how you're going to deal with this. You're going to come to one or two conclusions. One is, you're going to say, "Wow, I think I found a mis-priced bond. I get to buy this single A bond, but I get more yield than all the other single As. I get the yield of a BBB." Well, how great is that?
What you're basically saying is, "I think the rating agencies have properly characterized the risk and the market's just got the price wrong and I found it." We don't take that path, no shocker there, right? The other path you're going to take is, "Look, that thing's trading like a BBB because the market has figured out that it is a BBB and just got there faster than the rating agencies." So we've done all kinds of research. So every security was assigned an internal rating, every security and every portfolio based on how the market's treating it. In that example, we would give it a BBB.
So we could study how do these ones where we internally rated it BBB versus where the rating agency have called it a single A, what's its frequency of being downgraded? So these ones with these abnormally wide spreads, when you look out three months, six months, 12 months, they have about three times the frequency of being downgrades than the ones that are trading more in line with their peers. So on average, at any point in time, we would disagree with maybe 15 to 20% of the market value of the ratings out there, because we think the markets have priced in something different and we're going to put our faith in that market price more so than those rating agencies' assessments. On average, I think there's information in there, it's lagged. In the end, it's lagged.
Unreal. You got a 2016 paper that we talked about on this podcast, not that long ago, actually, where you took a lot of what we've talked about, about taking a variable maturity and a variable credit approach to fixed income. You built some simulated strategies using indexes where you varied the exposures based on everything that we've we've talked about. Now you used indexes. Do you think it would be feasible for a do-it-yourself investor, who, many of those are listening to this podcast? Would it be feasible for them to build a strategy like what you did in that paper using ETFs, that track the same indexes or similar indexes?
Well, they could try, but first, a couple of things I got to point out. You can't do this casually. It's going to be a full-time job for you. So you're going to go saying it's full-time. It's a full-time job for 25 of us in our group to do it. If you could do it with one, instead of our 25, that's pretty admirable. You can't do it casually. Not only do you have to keep track of it, to really do it, you would want to keep track of the U.S. Treasury curve, the agency curve, the AAA curve, the AA curve, single A curve, BBB, BB; then you want to do that on the 11 other curve. You want to do this every day to get a sense, so it's going to become a full-time job for you. Then, the next question is when you go buy those ETFs and you're going to be super active doing this, what about the implementation within the ETF? Now you have no control over that.
You're hoping that you're going to get index-like performance. You're going to get index-like transactions costs, which we think we can do better. So not only does it become a full-time job for you that you're really going to have to be monitoring all day every day, I would say at some point for all of us, you get to these make versus buy decision and that's an old engineering term and it probably makes a lot more sense to say, "Look, this is what these I have a whole group. This is what we do for a living." I think for many advisors, there's a lot of other important things they have to do in their day, service existing clients, try to go get new clients, all of these kinds of things and the investment strategies. I would think your chances are much better if you say, "I'm going to subcontract that out to someone else to run it for me."
So I guess the question that follows, for a do-it-yourself investor who doesn't have access to Dimensional, are they better off just holding a cap-weighted bonding index fund?
Well, let me put it this way. If my choices were between concentrated portfolios, making bets on their perceived mispricings or the index, I'd take the index for sure. If I'm one of those super low expense rates, if those are my only two choices, that's the one that I would go with. But if they do have access to Dimensional, I think our approach will lead to higher end returns and those index as well.
So for an investor who has the objective of getting the highest possible expected returns and can handle the volatility of an all stock portfolio, what might the arguments be to have some allocation to fixed income?
Well, if they can truly handle the volatility of an all equity portfolio, and I'll also say if their time horizon is properly set, we know that equities will have higher average returns than fixed in the long run. But when I say their time horizon is probably set, if you go a 10-year period where your equities underperformed T-bills, that can throw you for a loop. You got to be cool with that because there have been, and even longer than that, periods where the equity premium was negative. So you could say that, "I could handle the volatility," because you're thinking 18% 19, "Oh, maybe if gets bad, it spikes up to 30 or 40." I don't know, but you're thinking "Oh, yeah, yeah, I could handle that." You also have to have the right horizon. So I'm not opposed to some notion if someone says, "I want an all equity portfolio." You got to have the right horizon as well.
I want to ask one more time, David, because when we talked about that paper that you guys did using the indexes, we talked about that on our podcast in an episode, and people were super interested, oh, if they can do that with indexes, is there any way we could do this as do-it-yourself investors? If someone did a, I don't know, quarterly rebalancing or something like that, would that get anywhere close to being better?
See, in the end, look, this was for my research group. So if you were in that group and someone said, "Hey, we need to develop a simulation." Well, that's a really nice, handy way to develop a simulation. You're you get all these sub indices and they're blocks of securities and you put them in and out and you apply all of our rules. That's a great capability that we built, but I wouldn't go much beyond that. It's a good simulated portfolio, but to try to actually implement it that way through this whole array of ETFs, I think that'd be a nightmare, personally. I wouldn't recommend it.
You've been at Dimensional for a long time, so has Gene Fama. Presumably you've spent some time working with Gene. Do you have any good Gene Fama stories?
I remember my early days, we'd talk about something in fixed and they'd say, "Oh, run it by "Gene." I got to tell ya for a young fellow that was a terrifying experience at first. It's like my goal was to get him on the phone, ask my question and just get off as soon as I possibly can, because I was just so nervous. I think he sensed that. I call him, once I get on the phone, I think it was on a Tuesday, because I'm about to get off the phone. He said something like, "Well, did you watch that football game last night?"
I said, "Yeah, yeah. I did watch the game," and we started talking about the game. Next thing you know he's asking me about my kids and I'm asking, and we talked for like a half hour and I realized, "You know what? I don't have to be afraid." Anytime he wants to blow me out of the water with something technical, that could happen at will, but that's not what he wants to do. He really wants to help and help us do what we do and help me and so, I lost my fear for Fama after that call and I haven't had the fear of Fama ever since.
Great story.
So our final question, Dave. How do you define success in your life?
For me, ultimately, I think the measure is some long-term deep happiness, right? That's where you're being successful. It's not all money because I think you probably know people I think I know people that have a lot of money. They don't seem all that happy to me. I see people that don't have a lot of money and they seem really happy to me. So I would say, "Well, which one is actually more successful?" I think it has to do with that. That's sort of the measure. I think if you love what you do, that makes it a lot easier, but I think even more important than that is, if you love who you are that will ultimately, I think, lead to happiness. I think that's where it all starts. If you love who you are , and you've lived a successful life.
Well, it's pretty clear that you love what you do and that's been the experience every time I've seen you speak. I've never seen anyone more passionate about fixed income. This has been an amazing time with you, so thanks, Dave, very much.
Well, thank you. I'm happy to help and call me back anytime.
Links From Today’s Episode:
Rational Reminder on iTunes — https://itunes.apple.com/ca/podcast/the-rational-reminder-podcast/id1426530582.
Rational Reminder Website — https://rationalreminder.ca/
Shop Merch - https://shop.rationalreminder.ca/
Join the Community — https://community.rationalreminder.ca/
Follow us on Twitter — https://twitter.com/RationalRemind
Follow us on Instagram — @rationalreminder
Benjamin on Twitter — https://twitter.com/benjaminwfelix
Cameron on Twitter — https://twitter.com/CameronPassmore