Discover more from The MacroTourist
STOCK/BOND CORRELATION
The 'tourist examines what the flipping of stock/bond correlation from negative to positive might mean for markets
Summary
Leaning on the terrific work of Noah Weisberger from PGIM, we can identify three periods of stock/bond correlation.
The 1950-1965 period was one of negative stock/bond correlation, which was then followed by 1965-2000 where it flipped positive. Since 2000-2020, we experienced negative correlation. Determining whether we have returned to an era of positive correlation is one of the most important questions facing investors today.
To help determine the likely correlation environment, Noah identified both qualitative and quantitative factors which are now pointing to a new positive stock/bond correlation environment.
Determining the environment is crucial; “It’s a pretty unambiguous message which is that for a fixed-weight portfolio, a shift from a negative to a positive stock/bond correlation world leads to worse portfolio performance.”- Noah Weisberger
Portfolio managers, fiduciaries, investment advisors, and investors should take the time to understand the tradeoffs between increased volatility and maintaining portfolio returns in a positive stock/bond correlation environment. Even if there are no changes to be made, being prepared for the likely changed path of portfolio returns (more volatility) will ward off surprises which might lead to suboptimal shifts at inopportune moments.
Stock/bond correlation and the implications from a shift from negative to positive correlation
It’s easy to forecast that the stock/bond correlation might flip, but quantifying the factors that might drive that change, and even more importantly, what it might mean for markets, is a more arduous task. To help get it clear in my mind, I went hunting through finance papers. Lucky for me, I found Noah Weisberger, PhD. With the help of his team at PGIM, Noah has written a fantastic series of papers about the stock/bond correlation that I think are some of the best work out there on this subject.
His findings can be summed up in this 45-minute webinar titled: PORTFOLIO IMPLICATIONS OF A POSITIVE STOCK-BOND CORRELATION WORLD. For those interested in learning the nuances of Noah’s work, I strongly suggest you give it a full listen, but for everyone else, I have clipped the important parts and want to review his findings and add some ideas of my own.
Now, I know this might seem like a bit of a dry subject, but I believe the reversing of the bond/stock correlation will be one of the most important developments facing portfolio managers and investors this decade. It goes hand in hand with the reemergence of inflation, yet takes it to the next step by examining what it means for traditional portfolios.
All right, let’s jump in.
Noah: The salient feature of this picture (which kind of jumps out at you), is that on the one hand, we have experienced about 20 years (up until last year) of persistently negative correlation. This is the world many of us grew up in as professionals. But positive correlation is not unusual, and it can be very persistent. As we think about not only the historical pattern, but also the causes which could generate this historical pattern, the challenge is not to think with our long-term trend brain and not to think we our business-cycle frequency brain cause these regimes tend to be quite long lived.
I want to highlight that Noah is looking at stock/bond correlation in an extremely long-term fashion. He is using 5-year centered correlation based on monthly returns of the S&P 500 and the sovereign US 10-year Treasury index. According to his charts, stock/bond correlation is currently still negative, but we know that 2022 saw that relationship flip positive.
To better illustrate this recent change, I have created a chart of 1-year rolling correlation between daily returns of the S&P 500 and the Bloomberg 7-10 year Treasury Index.
For the period from 2007 to 2021, the stock/bond correlation was firmly negative. However, in 2022, it has gone positive, and is at the highest reading since 2000.
I believe we have entered a new era of positive stock/bond correlation, but let’s look at what Noah’s research says might drive such a shift.
Noah: A lot of the ultimate drivers of stock/bond correlation tend to relate to big fundamental features of the economic policy landscape - fiscal sustainability, monetary independence and supply versus demand side drivers. What seems to be the case is that positive stock/bond correlation tends to be associated with periods of unsustainable fiscal policy (or worries about the sustainability of fiscal policy), the knock-on effect that can have in terms of monetary policy where if there is a sustainability issue on the fiscal side that may lead to "less rules based" monetary policy and less independently driven monetary policy (where there is more incentive to let inflation run and help on the fiscal side), it tends to be associated where supply-side shifts are impacting the economy, and finally, when risks are being reweighted on a wholesale basis - it's not a relative risk story, but a risk across the board story.
What's really interesting is that we came up with these policies and big fundamental drivers at the outset of this research agenda in 2020, and it seems like we have ticked off a lot of the concerns we have been talking to clients about into 2021, 2022, and now 2023; what's going on with the debt and deficit side of the economy, what's going on with inflation in terms of supply-side drivers, how should the Fed respond (are they ahead or behind the curve), was inflationary transitory and they should wait or was inflation entrenched and they should have acted more quickly. Some of these big topics in the macroeconomic conversation end up having implications for the stock/bond correlation.
When you examine Noah’s checklist, almost every category is currently either in the “POSITIVE Stock/Bond Correlation” column, or has moved in that direction over the last couple of years.
Not content with a qualitative summary of what drives stock/bond correlation, Noah also quantified it.
Noah: There are three metrics which seem to drive stock/bond correlation (at least theoretically). One is the volatility of policy rates. The second is the correlation between economic growth and interest rates. And the third is the correlation between bond premia and risk premia.
And the equivalent statement of the stock/bond correlation being associated with debt sustainability concerns, monetary policy independence concerns, and supply-side drivers is the statement that positive stock/bond correlation tends to be associated with high volatility, high uncertainty in policy rates, negative correlation between economic growth and policy rates (policy rates are viewed as a surprise to the economy), and finally positive correlation between bond and stock risk premia (the wholesale re-rating of risk as opposed to the relative reweighting of risk preferences).
All right, this seems like a lot of theory, but it’s not as complicated as it sounds. Let’s tackle them one at a time.
The first being “volatility of policy rates”. Last month, I wrote a piece titled “UNDERSTANDING THE EPIC MOVES AT THE FRONT-END” where I highlighted how front-end volatility was as high as the 2008 crisis and the 1987 crash.
Obviously, the crazy moves we’ve seen at the front-end tick this box, so that’s the first check for the positive stock/bond correlation environment.
Next up is “negative correlation between economic growth and policy rates (policy rates are viewed as a surprise to the economy)”. For this, I will reference another paper by Noah titled “US Stock-Bond Correlation: What are the Macroeconomic Drivers?”:
I like to flip this backwards to explain it better. The period from the 2009-20 was a period with one of the lowest interest rates in modern financial history. Yet, economic growth was also some of the lowest reading of recent times. Economic growth was not negatively correlated to interest rates (lower rates did not spur more growth), and in fact, the opposite occurred. In Europe, you might even argue that negative rates caused moribund economic growth. The two series were positively correlated, which, according to Noah’s quantitative work, is more likely to create a positive stock/bond correlation environment.
Which then leaves us with our third factor; “positive correlation between bond and stock risk premia (the wholesale rerating of risk as opposed to the relative reweighting of risk preferences).” This seems recursive because if the risk premia of stocks and bonds are correlated, then it’s more likely that prices of stocks and bonds will also be correlated, but I understand the nuance. Noah is arguing that if “premia” - whether it be for term or for equity - are moving together, it’s more likely that the stock/bond correlation will be positive.
All right, we can debate whether stock/bond correlation has shifted into a positive regime or not, but for our next step, let’s examine the outcomes when it occurs.
Noah: What happens when correlation changes? Just to fix ideas, and from now until I toggle off these assumptions, the only capital market assumption we will play with is correlation itself. A lot of people link correlation to the broader capital market landscape, but let’s put that aside for a moment and assume that stocks and bonds return their historical averages over the last 50 years, so we’re going to play around with what correlation goes with that set of returns and volatilities.
If we start off in a world of 12% stock returns and 7% bond returns, 15% stock vol and 8% bond vol, and we assume a correlation of negative 0.25 (which in a negative correlation regime isn’t an outlandish number - it seems rather run of the mill), and we look at a 60/40 portfolio…
… you can see that at point A on that light blue efficient frontier, you have an expected return of about 10% and a volatility of 9%.
Now, again - holding all other capital markets assumptions fixed - we’re now going to change the correlation regime from negative 0.25 to positive 0.25. You see how the efficient frontier is now going to shrink in [editor’s note: the black line].
At the very ends, it’s the same. If you have no stocks or no bonds, you’re back to the same portfolio, but in the middle area, there is a shrinking of the set of portfolios that are achievable.
You can think about this choice as one of two: if you are a fixed-weight portfolio and you don’t want to change that allocation, then you’re going to need to shift to the same expected return, but with a higher volatility (that’s Portfolio C).
Alternatively, if you have a very strict volatility or risk budget, you’re going to have to accommodate the higher volatility landscape by lowering your allocation to the risk asset and by accepting a lower return [editor’s note: Portfolio B].
This sets up the framework by which the channel you need to think about portfolio performance is probably in volatility and then we’ll get to a moment where you’ll determine where do you move (Portfolio B, Portfolio C, somewhere in between).
Another way to think about this is as follows; when stock/bond correlation is negative, there is a dampening effect to the volatility of the portfolio because of the fact that when one asset is rising, the other one is more likely to be declining. In a positive stock/bond correlation environment, the moves are more likely to be amplified because the two assets increasingly will move together. Therefore, as stock/bond correlation increases, the volatility of the 60/40 portfolio increases, and an investor faces a choice of accepting that higher volatility or reducing the risk asset at the expense of returns.
But how important is this stock/bond correlation to long-term returns?
Noah: For CIOs, we’re not thinking about one period, we’re thinking about a longer-term like a 10-year performance, and really - it’s not period, by period, but the terminal distribution of wealth outcomes at the end of that 10-year period.
So now we are going to simulate stock and bond returns, and we’re going to use the actual distribution of stocks, so we’re going to allow fat-tails and other non-normality returns to creep in, we’re going to keep our assumptions about annualized stock and bond returns and volatilities, we’re going to focus on 60/40 for now, and we’re going to ask ourselves; what happens when we go from a negative 0.25 correlation world to a positive 0.25.
You see the pickup in portfolio volatility. If you follow that middle set of blocks throughout these charts, what we are pointing out here is a very clear message; when correlation shifts from negative to positive, long-term risk-adjusted portfolio performance suffers. You have a wider distribution of terminal wealth, you have lower risk-adjusted terminal expected wealth, you have deeper drawdowns, worse tail events, and a higher probability of ending the 10-year period with less money than you started. It’s a pretty unambiguous message which is that for a fixed-weight portfolio, a shift from a negative to a positive stock/bond correlation world, leads to worse portfolio performance.
Not much to add as the last sentence says it all.
I believe the table has been set for a sustained period of positive stock/bond correlation. All of the factors outlined by Noah appear to be indicating the end of the 2000-2021 period of negative stock/bond correlation.
If we have entered a period of positive stock/bond correlation, it’s important that portfolio managers, fiduciaries, investment advisors, and investors prepare themselves for either increased volatility or decreased returns for their traditional portfolios. Some might argue that they don’t invest in the 60/40 portfolio, so this doesn’t apply to them. I would take the other side of that claim. Whether you invest in 60/40 portfolio directly or not, the concepts of risk asset/safe fixed-income asset correlation are likely embedded into your portfolio assumptions. It’s extremely difficult to escape this reality. And even if you have somehow evaded that inevitability, the increased volatility of most investors’ portfolios could likely see a mass selling down of all assets at the wrong time, so it’s still important to understand the changes to the financial landscape that might occur if stock/bond correlation flips positive over the long-run. These are obviously extremely drawn-out trends, so it’s not something you need to run out tomorrow to fix. But I think that Noah’s work is instrumental in understanding the causes and consequences of changing stock/bond correlation. I’d like to thank him for all his hard work, and in the kindness of sharing all these findings.
Next post, we figure out what we might do to help our portfolios in a positive stock/bond correlation environment.
Thanks for reading,
Kevin Muir
the MacroTourist