REPLAY: PLSA Investment Conference 2022
David Hunt Speaks at PLSA Investment Conference 2022
Noah Weisberger, Managing Director at the PGIM Institutional Advisory & Solutions group, joined Investment Magazine’s Head of Institutional Content, Alex Proimos, for a discussion on today’s market dynamics fueled by fiscal stimulus and monetary policy, and shares insights on how investors may want to allocate risk across their portfolios under these conditions.
>> I've got the great pleasure of being joined by Noah Weisberger, who is the Managing Director of Institutional Advisory and Solutions at PGIM. Noah, welcome.
>> Hey, Alex. Thank you very much for having me.
>> Look, it's a great time to have your session, particularly after we've just had three CIOs really talking about some of the key issues that they face. One of the other really big issues that they face is this time in the market where, whether bonds and stocks moving in this negatively correlated way still holds forward. Maybe if you could give a bit of a backdrop to the research that you've been doing around sort of correlations between fixed income and equities over the past sort 100 years, it'd be really useful.
>> Yeah, absolutely. That's a really great context to put the question that we're asking and it really is maybe it's coming a bit to a head now. But as you know, for the lifetime of most of the CIOs that we speak with stock on correlation, the correlation between stock returns and bond returns has been persistently negative say, since around 2000. So, that's a 20-year period of a pretty persistent negative correlation with a lot of really important implications. So, if you're building a portfolio and the building blocks are stocks and bonds, and we're looking in the US currently. So, if you're building blocks are say in S&P 500 and the treasury, there's a natural hedge there from that negative correlation. And what that means is that the portfolio for any given level of risk has a little bit -- for any given level of expected return, has a little bit risk -- a little bit less risk than it would otherwise given that negative correlation, that natural hedge. But what we're trying to remind investors is that the relationship between stock and bond returns is not immutable. These regimes are very long lived. But they do change over time. And so, prior to 2000, there's almost a 35-year period 1965 to 2000. And think about all the things that changed in that period. Well, one thing that didn't change is that stock-bond correlation was fairly persistently positive all throughout. And then if you roll back even further in time post World War II until about 1965, stock-bond correlation was negative the way it is now. So again, very, very long live regimes but regimes that do change, and losing that natural hedge, losing that negative correlation, and moving into a positive correlation regime, we think that can have some impacts on the portfolio and something that's worth focusing on.
>> Well, let's go into that whole evolution from negative to positive, you know. Particularly now, we've run a 40-year decrease in interest rates that was benefiting both fixed income and equities. Now, we're seeing that change, you know. Is that what we should expect now for the next 10 or 15 years? Or we'd be moving through different regimes as we come out of this?
>> So, I think there are a couple of different things that are at play that we would highlight. And I think the way we've come at this is almost from first principle. So, we think about what it is that determines the price of a stock. What it is that determines the price of a bond? What are the commonalities between those two assets?And then, what drives the correlation of returns? And in that context, you know, we think there's a macroeconomic decomposition that can be done. That relates a little beyond just sort of the level of interest rates and the trends we've seen, all those long multi-decade trends that we've seen. And relating to things more like, you know, the volatility of interest rates as opposed to their level. Maybe also, what you're alluding to, I think, is an important component there is not just the level of interest rates, but the risk premia associated with both bonds and stocks. And so, when those are moving together, when the market is re-rating both bonds and stocks that tends to impart a positive correlation, for over long periods of time, those moves are dominated by negative correlation when relative prices, relative risk prices of bonds and stocks are in flux. And then finally, probably the component, the macro component, if you will, that I think is the most interesting to think about from some -- in some respects is the correlation or the co-movement of economic growth and interest rates. So our economic growth and interest rates moving together. And that has some real policy underpinnings to it, versus periods of time where they're moving apart, where interest rates go up, and growth goes down, interest rates go down and economic growth goes up. That's a very different policy setting. So, as we try to unwrap sort of this puzzle, we think stock-bond correlation can be broken down into macro bits and pieces. But those bits and pieces themselves, really, we think, reflect policy settings both in the monetary and on the fiscal side and also, the balance of supply and demand drivers at play at any point in time. So, those would be the building blocks. And the question is, you know, which were active in the different historical periods and what might change on a forward basis?
>> How easy is it for you to analyze what are the key driving factors for this correlation import whether it's around fiscal policy, for example? We see more and more of that of late. But historically, we saw a lot more movement on monetary policy whether it was interest rate change, liquidity. How much can you basically pull out those factors to understand this correlation?
>> Yeah, I mean, those are the ones. I mean I think it's -- we're focused on the big macro components. So obviously, there are lots of other stories that can float around, you know, some to do with market structure or higher frequency issues. But I think pulling apart these basic macro components, I think can be done. We think the data historically are pretty dispositive in the sense that they do seem to, if not dictate, at least coincide with these long regimes. And so, previous periods of positive stock-bond correlation tend to be associated with higher interest rate volatility, with negative correlation between economic growth and interest rates. And with trending positive co-moving risk premia, and sort of flipping that all around. If we look at the 2000 to the 2020 period of negative correlation, the opposite is true. That there's a period of relatively lower volatility, interest rate volatility. It's a period of positive co-movement between economic growth and interest rates. And it's a period of negative co-movement between the risk premia. We think all of those are the factors that support positive or negative stock-bond correlation. And then mapping that to policy, we think has a lot to do with thinking about not only sort of the policy buckets on their own but critically how they interact with each other. And so, to sort of put a bit of a wrapper around it, our analysis, the point we get to is that if we were to see a shift from negative to positive stock-bond correlation, so going from the current negative regime to one where stocks and bonds were no longer hedges for each other, we think some of the macro policy preconditions would be something along the lines of monetary policy that's a lot less rules based than it's been in the past. More discretionary monetary policy we think would be less sustainable or perceptions that fiscal policy is less sustainable. And when I say both of those two things, immediately, what should come to mind is that the benefits, the potential benefits, and the incentives for policy coordination, something that we haven't seen in the US for decades and decades, also is present. And we think that loss of fed independence or perceptions of loss fed independence, and of coordinated policy, which just reinforced some of these changes in the way economic growth and interest rates interact with each other with the volatility of interest rates and ultimately, for an investor and for a CIO impacting stock-bond correlation. I think the other part that would play into this is this will probably lead to a re-rating of risk in both markets, which would then reinforce the positive co-movement of those two markets which would then reinforce the re-rating. With this hedge absent you'll probably need to think differently about risk pricing of these two assets. And the final thing I'll say about where we are currently, you know, if we think about how much we've seen a compression in risk price in both markets, it's hard to envision if those risk prices are going to move together, to trend further lower. The positive co-movement story would have to be some re-rating higher risk in both markets. That's not to say it's going to happen. But it is to say, my guess would be if we're going to see a positive comovement between equity and bond risk prices, it would have to be higher, not lower.
>> So, if you put yourself in the position of a CIO and they're trying to look at the regime today or where we're moving into what particularly could they look back through history and say, "Well, this is what we're referring to?" Is there something that they can look at that would give them a bit of clarification of how this could potentially play out?
>> Absolutely. And I think, you know, and I'll put kind of the risks on the table even more explicitly. I think the worry is that we've seen, at least in the US, we've seen a big buildup in debt to GDP. We've seen deficits deepening. We haven't seen the market react too much to that. And so, a lot of people look at the gap between interest rates and growth as a measure of how sustainable a fiscal position is. We haven't really seen that move too much. But the movements in debt to GDP and debts to sort of harken back to maybe the 60s or the 70s, the shift in fed policy, the explicit shift in fed policies, interests, if they could pull this off, then it would be a new way of targeting inflation. They've been incredibly transparent and credible up until this point. But I think the risk is that as we move towards a new policy regime with this long-term average inflation target, the market isn't quite sure how that's going to be run. And I think the analogies to the movement from the 60s into the 70s is a good one in the sense that late in the 60s and early in the 70s, I think fed policymakers felt they had built up a great deal of credibility. You know, inflation was averaging sub 2% in the mid to late 60s, a level that took us almost 40, 50 years to get back to currently. And so, I think credibility is hard won and usually hard loss. But once it's lost, it's very hard to win it back again. So, we have a little bit of uncertainty about exactly what the fed means and is going to do. We have a little bit of uncertainty about how sustainable the current fiscal stance is. And I think that sort of creates the incentives for certainly something that we've seen talked about more in my professional lifetime now than any time in the past, which is the incentives for some coordination, either explicit or otherwise. And, you know, some people look at quantitative easing in the expansion of the fed balance sheet as in essence steps in this direction. And part of that is a playbook, right? When we have incredibly low interest rates, negative interest rate -- sorry, zero lower bound interest rates and arguably a liquidity trap, the right thing to do is to use fiscal policy to generate demand growth. And at the same time as the existing burden of debt grows with these increased deficits, then the incentives to use other policies on the monetary side also grows as well. And so, I think there's a lot of uncertainty around the path forward and I think those will be some of the things, the conceptual things to look for. I think there are some more pedestrian indicators we could talk about as well. But at least conceptually, I think that's the debate, where the policy debate is. And I think it has real concrete implications for CIOs in terms of these two core assets in the portfolio and how they're likely to behave one against another with the implication, of course, in terms of correlation and the implicit head that one asset has than the other in the portfolio.
>> Is there a need then for a CIO to almost get a feel or in terms of your modeling, to get a feel of sentiment and liquidity and I guess, financial conditions, as part of trying to understand this correlation?
>> Yes. So, I think where that really manifests itself is how you think about the interaction of economic growth and interest rates. And when the fed is operating in a rules-based manner it's almost responsive to inflation, inflation expectations and growth. And so, in that sense, as economic growth increases, fed follows suit and interest rates do as well. And that positive growth rates, interest rate correlation, is what we think leads to a negative equity bond return correlation. The flipside would be is if the fed is trying to engineer something or act in a way that's not consistent with a rule-based policy. And I think that will lead to the other set of outcomes in terms of positive correlation that we were discussing. So, I think, thinking about what the fed is doing, what they're likely to do, and what it looks like relative to rules is very important. There's a very kind of easy way to look at this, which is to look at the actual effective fed funds rate relative to a rule-based rate. The difficulty, of course, is that, you know, the fed funds rate has been at the zero lower bound so, we've had to do all these other extraordinary policy tools being used. And those lead I think to a little bit of confusion as to, you know, how transparent can the fed be when they're using these sort of nontraditional tools, which actually become traditional, if you will. So, I think that makes it a little bit more difficult to use some of the 1992, 2000 era toolkit. But those would be some of the things that I would focus on a lot. And on the fiscal side, I think it's obviously the path of debt to GDP deficits but also, critically, this difference between interest rates and growth. And a lot of academics and practitioners as well have opined that as long as the economy can grow faster than the cost of debt, ultimately, if deficit stabilized, the debt can be viewed as sustainable. But looking back over time, that's a very reflexive relationship, which is as soon as the market loses confidence in sustainability, interest rates will move, which will crowd out and slow growth. And so, you can look sustainable until a period of time at a breaking point, you know. When those come, you know, that's not science. But I think these are the important things that we think should be on CIOs' dashboards not only because stock-bond correlations is interesting to me, you know, artifact, but because it's a really important building block of a lot of portfolio construction assumptions.
>> You mentioned there about the engineering or the engineering of the market by central banks, obviously, and then fiscal policy helping. Is there a concern that some of these engineering that's been going on to help the economy is actually suppressing volatility, and hence, maybe not seeing the ultimate correlation between stocks and bonds that should be there because of the work that's being done by these different parties?
>> Yeah, that's a really interesting question. And I would layer another observation on top of that, which is we've just been through or we continue to go through a really massive cyclical re-rating so, like the -- I don't want to use too many superlatives, but a quite large risk offer has gone a period of time with the fed committing themselves to low rates. And so, I think the endgame here is twofold which is we've seen a very, you know, a re-doubling of this negative correlation throughout the big market selloff late last year in the rally that we're still enjoying now with bonds. Losing ground with equities, gaining ground rather quickly. We've seen a very strong positive correlation between economic growth and interest rates meaning, as economic growth slowed we saw interest rates fall dramatically. And as economic growth has recovered, we see interest rates start to rise. All of that reinforces the negative stock-bond correlation that we've been living with for a while. The feds' commitment to low interest rates, suppresses rate volatility and all of that, again, reinforces negative stock-bond correlation. Is it artificial or not? I don't think we can know that yet. But I think this lower for longer commitment, if it's really just a function of the fact that economic growth is still soft, you'll go back to the conversations we were having three or four or five years ago about secular stagnation. If that's the world that this big risk off has gone and sort of hiding and what we return to, low rates is not suppressing volatility, it's reflecting the true tepid, underlying economic growth. But if the Fed is suppressing rate volatility, keeping rates low because they have other objectives in terms of supporting, you know, debt payments, or things like that then, I think we could go to a very different world. I think this conversation has been pushed off by a year or maybe more because of the big cyclical swings, which I think will reinforce negative stock-bond correlation for a little while, at least. But from here, you know, we could envision a return to, you know, the current state of affairs in terms of negative correlations as I was trying to outline. We could see a new paradigm that would be kind of in a not very benign environment where fed credibility and fed independence is questioned along with fiscal sustainability. But there also could be a middle path, you know. And I think back to the 1990s, where we had exceptionally strong economic growth but supported by some unexpected supply-side shifts and productivity gains. And maybe that would support a lower for longer low rates vol world but maybe with a different, you know, set of drivers for the overall, you know, macro economy.
>> Now, I've got a whole heap of questions here. So, we need to get through them as best we can. So short answers, if possible. Is the argument predicated on the US dollar being the world's leading currency and would it -- if it loses its status, what impact would that have on correlation?
>> So first, I would say short answers are just unreasonable but I'll try my best. So, is it predicated on the dollar? I think it's predicated on some combination of dollar and bond as sort of the riskless asset. And obviously, we could put a little bit more risk premium into that riskless asset and that's what we're talking about. But I think we had a change in regime and the US dollar and the US bond market were subject to the same global flows. I think maybe we'd be in a little bit of a different position. One way we wanted to get this question and I'll take 30 more seconds is to look at global bond and stock markets in this context and see, is the US one of a set of markets that are kind of all behaving the same way? Or is the US the leader? And then, I think that question would carry a little bit more weight.
>> We actually had a similar question about that in terms of the correlations globally, in different regions and different policies. But I'll go back to another question, which is around, many investors start with expected returns, then volatility and then they look at correlation. Is it best for investors or CIOs to turn this around?
>> That's a great question. And I think I would answer to is one, all else equal, a change of correlation can be very damaging to a portfolio, for a CIO to make a choice between, you know, maintaining an expected return target and suffering with higher volatility, or lowering the expected return, lowering the risk allocation but maintaining the risk budget. Historically, at least in the US, what's really interesting, surprising to me in fact was that the 60/40 portfolio let's say, did not perform markedly better ex-post during periods of positive correlation. I think it's because correlation regimes tend to be really, really long. And so, I want to give a maybe not a very satisfying answer, but I think that the empirics suggest that most of the time the focus remain on first means then vols then correlation because the correlation regime is so long and stable. I think there are periods of time when the risk of a shift in regime are heightened. When that correlation as a very big estimate could shave off, you know, a 10th of a point to a Sharpe Ratio or more than that. And in the competitive landscape when things are in flux, getting that extra little bit of Sharpe Ratio I think is very impactful. On an average month, day, week or year, no, I don't think correlation is the first input. But I think being attuned to regime changes is very important.
>> I'll come at you with a final question and on this correlation piece. It comes through from anonymous, but that's okay. All the funds typically use stationary correlation matrices. Should they be moving to a multi-state Markov chain approach for different correlations and different regimes?
>> So, this is the short answer. Yes. You know, we haven't done the entirety of that work. It's definitely on a whiteboard. I think a conditional correlation matrix is a really important way to think about it. I think the hard part is, as I was trying to outline is that oftentimes correlation doesn't change very frequently. And being off a little bit on correlation I think is really -- I don't know even know if it's second order, it might be third order but when it shifts -- and the other thing I would say, just to add, given that this question uses the word Markov chain, which I was told never to use in a public forum. But now, that the cat is out of the bag, the relationship between let's say, a metric of portfolio performance like Sharpe Ratio and correlation, so, if you think about that relationship for a given change in correlation, the change in Sharpe Ratio is highly nonlinear. It depends on lots of other states of the market. It happens to be that when correlation is starting off negative, when volatility is low and expected returns are high, a change in correlations have the biggest impact on risk reward metrics. So, you know, we're at risk. I think we need to be vigilant given the state of the US policy debate. And I think we need to be vigilant given the state of markets that we could be more susceptible than other points in the market dynamic given kind of the math of correlation Sharpe Ratio.
>> Well, we've really opened up Pandora's Box with the amount of questions that came through for this session, which is fantastic. And I guess for the audience at home, there will be a follow up podcast on the market narrative series to go into this in some more of I guess more depth and understanding of how to use this as a CIO later on. So, thank you very much for your time today, Noah.
>> It was a real pleasure, Alex. I enjoyed being with you.
>> Thank you.