Jeffrey Knight, CFA
Global Head of Investment Solutions and Head of Global Asset Allocation
Columbia Threadneedle Investments
P&I: What are the big ideas behind Columbia Threadneedle's adaptive risk allocation (CARA) strategy?
Jeffrey Knight: It's very easy, using conventional asset allocation techniques, to allow your portfolio outcomes to be dominated by a concentrated influence like the stock market. Many investors understand that even a portfolio that appears broadly diversified from a capital perspective — say, a traditional 60% equity/40% bond mix — has a very unbalanced risk allocation, with nearly 90% of portfolio risk driven by the performance of equities.
So big idea No. 1 is that a better approach to allocating assets is to think about how your overall return is influenced by sensitivities to the individual component parts of a portfolio, and to try to keep those sensitivities in balance — not let any asset class have an outsized and unintended effect on portfolio returns. That is the basis of many traditional risk parity strategies, and most of the time, that is an advantageous way to organize your asset allocation.
Big idea No. 2 recognizes that there are certain market environments when risk-balanced approaches to asset allocation are not optimal. One clear and common example being an environment in which the risk and return profiles of the constituent portfolio building blocks (stocks, bonds, etc.) are lopsided. When some assets look strong and other assets look weak, it makes better sense to emphasize those investments that look strong. Another example is an environment in which all the portfolio components appear equally unattractive and could decline simultaneously, like during the taper tantrum in the summer of 2013. In such a circumstance, diversification, either capital- or risk-based, doesn't help you very much.
Big idea No. 3 synthesizes these first two observations, and posits that evidence-based research can provide tools for anticipating those conditions when balanced risk investing may be sub-optimal. When those exceptional conditions are flagged, we don't merely tweak our existing portfolio, we shift completely to a different asset allocation more suited to the new environment.
P&I: How do you make sure that you are moving the levers at the right time so you are not behind the curve when those changes take place?
Jeffrey Knight: There is no “sure” in investing. I am perfectly happy not to force changes into the asset allocation, so we wait for exceptional circumstances before we abandon our neutral, balanced risk approach. We use common sense stock and bond indicators to identify atypical conditions. These indicators move around all the time, but we only take action when they depart in a meaningful way from their normal range. It takes more than one basis point or a one-day trigger to prompt a shift in the allocation. That patience and discipline is intended to keep us from reallocating either prematurely or unnecessarily.
P&I: Looking at today's economic and market environments in the context of those big ideas, what are the bond and equity markets signaling?
Jeffrey Knight: It helps to look at the current environment in the context of where we have been recently. For much of 2016 up until the U.S. presidential election in November, almost every asset class delivered positive returns. This was a wonderful time to employ a risk-balanced approach. However, things changed after the election; investors considered a shift away from policy gridlock and experimental monetary policy, and we saw a profound shift in context for investors — equity markets have done very well, but bond markets have done very poorly. It has been a suboptimal period for balanced asset allocation; you would have been better off letting stocks dominate your portfolio.
While our methodology for determining market state is blind to specific external circumstances like a presidential election, the simple metrics that we track to determine bond market and stock market conditions did reveal an elevated probability of a lopsided market environment. Our bond market state going into November 2016 was negative, meaning that interest rates were too low compared with prevailing conditions. This condition suggested that, if risk assets were to perform well, bonds would likely re-price and underperform sharply. At the same time, our equity market indicators registered as positive, owing to an unusually favorable combination of our equity indicators on a global basis. We call the market state corresponding to negative bond market conditions and positive stock market conditions our highly bullish market state. Our policy portfolio for the highly bullish market state emphasizes risk assets like equities and dramatically reduces interest rate sensitivity. While it felt uncomfortable at the time, the strict adherence to the policy portfolio determined by the market state turned out to be very beneficial. Further, most months since the election have mapped to the highly bullish market state.
More recently, we have observed a different reading from our indicators. It's too early to say how it will play out, but the market environment appears to be on the brink of a change from what we have seen over the last six months.
P&I: What are you observing?
Jeffrey Knight: I still think bond yields are chasing fundamentals higher. We have seen, for example, inflation in the U.S. go from less than one percent last summer to closer to three percent. Bond yields are up significantly, but the markets may not have adequately recalibrated to that new level of inflation.
Bond repricing could go on for some time, and if we find ourselves in a world of accelerating growth and inflation, we could see yields go higher for months. I do not think that scenario is out of the question based on policy expectations from the new administration or from the stronger recent economic data that we have seen.
As far as equities are concerned, the market has been in a pretty steady rally since the election, and we are watching the indicators for a shift in the reading that has been in place over the past six months. One of these days, we are going to see a change, and it is certainly not out of the question that we could enter a period of sustained difficulty for equities.
If equities struggle while bond yields are priced in such a demanding way, it could spell trouble for balanced risk strategies. For such portfolios, a lot of people fret that rising rates are going to cause all the damage. I disagree. If stocks are rallying while rates are rising, you have some protection due simply to multi-asset exposure. The problem scenario is one where stocks and bonds decline together for an extended period, and that scenario cannot be ruled out. At the moment, the indicators are not definitively aligned with this scenario, and we are mapped to a neutral allocation.
P&I: What would the policy portfolios look like in these scenarios?
Jeffrey Knight: One of the other big ideas underlying our approach is an observation that even if we recognize a need to reduce risk, we may not respond with sufficient force if left to our own human judgement. That is why we utilize pre-specified policy portfolios corresponding to each of our four market states. If our monthly indicators were to shift from a highly bullish state to a capital preservation state, our risk exposure to equities would be cut from nearly 70% to less than 20%.
P&I: You have said that we are moving away from a world defined by policy gridlock and experimental monetary stimulus. Have you noticed or identified any changes in how that shift is playing out since the election or the inauguration?
Jeffrey Knight: Given the Republican sweep in the U.S., the possibility of economic policy drawing upon a broader suite of tools is a tantalizing one — tax reduction, deregulation, or fiscal spending are all possibilities. We have not seen any of that for a very long time so the burden of economic resuscitation fell on monetary policy, which led us to zero interest rates and quantitative easing, and similar interventionist measures.
What we expect is a meaningful transition to a more full-fledged economic policy, with less reliance on monetary policy. In the U.S. we are seeing some evidence that we are going in that direction. There have been conversations about fiscal spending and tax reform, and on the monetary side, we have seen two rate hikes. I think the markets have been pricing in and anticipating a shift away from fiscal gridlock and monetary stimulus to more explicit and multi-dimensional growth-oriented economic policy.
The risk is that we don’t get what we are expecting. But it’s just too early to say one way or the other.
P&I: How might that shift and those expectations play out in the adaptive risk allocation model?
Jeffrey Knight: We still think interest rates are too low relative to fundamental conditions. That could resolve itself in an environment where risk assets continue to rally, and the growth-oriented market pattern continues. In that case, we would likely see the CARA framework signal bullish, with a policy portfolio tilted towards risk assets.
However, bond market disequilibrium could also reconcile by the fundamentals moving to where the bond yields are, instead of the bond yields moving to where the fundamentals are. That would probably happen in the context of turbulence in risk assets like equities, and would be more compatible with what we call our capital preservation market state. In this scenario, the name of the game is to keep your portfolio safe and allocated in a way that is very resilient. In that environment we would expect to see a drawdown in risk assets, and very little diversification benefit to be gained from bonds, because they are already mispriced.
Between those two, we have our neutral state, which employs a risk-balanced approach, and seeks to deliver stable returns by just being balanced across stocks and bonds.
At the moment, we find ourselves comfortably residing in the neutral market state. But our methodology is suggesting that we are nearing a pretty significant inflection point, where staying centered and balanced is likely to be suboptimal. It is just not clear yet, whether the most efficient portfolio is one tilted towards, or away, from risk. ■