22.07.2024, 17:46 Uhr
In der Auseinandersetzung um die Besetzung des VR-Präsidiums bei Logitech hat sich der Firmengründer Daniel Borel durchgesetzt. Auf seinen Antrag muss VR-Mitglied Guy Gecht für das Amt nominiert werden.
Hellish Choices: What's An Asset Owner To Do?
Ben Inker, Portfolio Manager at GMO, comments about the actual low interest rate environment and the high ratings of almost all asset classes. Therefore, lower returns than historically measured are guaranteed in the future. Will the ratings return back to normal or stay at an increased level?
Over the past few years, my colleague James Montier and I have written extensively on the possibility that there has been a permanent shift in the investment landscape. The investment landscape today is an unprecedented one, where we believe it is not so much that asset prices look mispriced relative to one another (although some of that is going on) but that almost all asset classes are priced at valuations that seem to guarantee returns lower than history. Our standard forecasting approach assumes that this situation will gradually dissipate, such that seven years from now valuations will be back to historical norms. James calls this scenario Purgatory because it means a finite period of pain, followed by a return to better conditions for investors. An alternative possibility, which James refers to as Hell, is that valuations have permanently shifted higher, leaving nearer-term returns to asset classes somewhat better than our standard methodology would suggest, but at the expense of lower long-term returns. By now some of our clients are probably thoroughly sick of hearing about the topic, but this piece is going to delve into it yet again, because the question of whether we are in Purgatory or Hell is a crucial one, not only for its implications for what portfolio is the right one for an investor to hold at the moment, but also for the institutional choices investors have to make that go well beyond simple asset allocation. In his letter this quarter, Jeremy Grantham is exploring a slightly different version of the Hell scenario. His version of Hell is driven more by weaknesses in the arbitrage that should force asset prices back to equilibrium rather than changes to discount rates, but it has similar implications for investors and institutions.
The distinction between Purgatory and Hell is an important one, because each scenario creates different challenges for investors. In short, Purgatory creates an important investment dilemma today, as the optimal portfolio looks strikingly different from traditional portfolios. If we are in Hell, by contrast, traditional portfolios are not obviously wrong today, but the basic assumptions investors make about their sustainable spending rates are dangerously incorrect. These twin issues the implications for todays portfolios and the implications for institutional decisions into the future are crucial for us in the Asset Allocation group at GMO. And they are the reason why most of the investment conversations I have had with my colleague John Thorndike over the past few months have either been on this topic or have been impacted by the dilemma of which scenario to assume. The rest of this paper is largely a summary of our discussions on the topic, which we recently presented at our annual client conference.
Why all the discussion of Hell?
There are a couple of important points to make about why we spend so much time discussing the Hell scenario today when we have not done so in the past. The first is that we would likely not be discussing it at all if it did not seem to be priced into the market. If asset prices today were generally clustered around normal valuation levels, we would spend very little time concerning ourselves with the possibility that all assets would simultaneously rise in price and valuation to a new high plateau, which would, in the future, be considered normal. In other words, we run the risk here of making the same mistake that we have accused plenty of other investors of making in the past, trying to justify the price level of an overvalued market with those ever dangerous words This time is different.
The second is that unlike many points in history, the change to the expected returns to assets involved in going from a Purgatory scenario to a Hell scenario has very important implications for the correct portfolio to hold today. Even had you known in 2000 or 2003 or 2007 whether the future was going to evolve consistent with the Purgatory scenario versus the Hell scenario, you would not have needed to run a particularly different portfolio. Todays portfolio is much more profoundly different depending on which you think is correct. And finally, Hell goes beyond being merely an investment problem into the realm of being an institutional problem, with implications for foundations, endowments, pension funds, and individual savers that are much broader than the simple question of what portfolio is the right one to run today.
Purgatory and Hell, the seven-year view
Our standard assumptions for long-term asset class returns are that equities should deliver 5.5-6% above inflation in the long run, bonds 2.5-3% above inflation, and cash 1-1.5% above inflation. These assumptions are broadly consistent with the long-term returns to each of these assets, and embody what we feel are appropriate premia for bonds and stocks above cash, given the risks that each impose on holders. The assumptions for Hell are 1.25% lower equilibrium returns across the board. In other words, in Hell the equilibrium return on cash is 0% real and risk premia are otherwise left unchanged. The implications for our asset class expected returns over the next seven years are shown in Exhibit 1.
As you can see, for all assets other than cash, expected returns are higher in Hell than in Purgatory. This is because, other than cash, these are all reasonably long-duration assets, and gains made by not having to fall to the lower valuations of Purgatory outweigh the lower income generated by the fact that valuations are higher on average over the period. An assumption of higher ending valuations will always make for higher expected returns over a seven-year period for stocks and bonds. But for much of history, this wouldnt necessarily have affected the correct portfolio to hold. Exhibit 2 shows the result from a simple optimization given our standard 7-year forecasts as of June 2000 and the result if we adjusted the forecasts for the Hell terminal assumptions.
While the expected returns are different for the two scenarios, they have no impact on the desired portfolio, which winds up buying as much emerging equity, emerging debt, and TIPS as it is allowed, with the rest in US government bonds. In 2007, the impact of Purgatory versus Hell is slightly larger, as shown in Exhibit 3.
The basic difference is how much you want to have in quality stocks 10% versus 30%. Interestingly, despite a 20% swing in the amount in equities, the expected volatility and duration of the portfolios winds up quite similar, and the expected returns are exactly the same in Purgatory and only differ by 0.4% annualized in Hell. The duration I am calculating for these portfolios is the version I used in the Q2 2016 Letter The Duration Connection. It takes the basic concept of duration sensitivity of the price of an asset to changes in the discount rate and applies it to all assets, not just fixed income assets. The specific durations I am using are identical to the ones I used in that letter.If we do the same experiment today, we get two quite different portfolios, as you can see in Exhibit 4.
The Hell portfolio has 58% in risky assets versus 36% for the Purgatory portfolio. It has almost twice the expected volatility and close to three times the duration. And while the two portfolios have similar expected returns if Purgatory turns out to be the correct forecast, the difference between the two in Hell is 1.3% per year for seven years. Another way to look at it is if you pick the Purgatory portfolio and it turns out we are in Hell, you will only break even with a more traditional 60% stock/40% bond portfolio. If you run the Hell portfolio and it turns out we are in Purgatory, you are taking almost twice the volatility and expected loss in a depression, and almost three times the portfolio duration for no benefit in expected return.
What can an investor do to try to solve this problem? At first blush, risk parity seems like it might be a solution. After all, risk parity avoids the task of forecasting in the first place. We dont think it helps, though. One reason is that the fact that a strategy does not involve forecasting doesnt mean that expected returns for assets dont impact the expected return of the strategy. And today we believe that risk parity has two problems on that front. The first is that global government bonds have a negative expected return versus cash in either Purgatory or Hell. The second is that the duration of many risk parity portfolios is extremely long today. Part of the reason this has occurred is because the duration of government bonds has lengthened as yields have fallen. But the issue has been amplified by the fact that volatilities have been low and correlations between stocks and bonds quite negative, which means risk parity portfolios that key off of portfolio volatility targets have been levering up. This long duration means that such risk parity portfolios are particularly vulnerable to any rise in the general level of discount rates for asset classes, and their vulnerability is greater in both absolute terms and relative to a 60% stock/40% bond portfolio than has generally been the case over the last 40 years. Exhibit 5 shows the duration of a simple stock/bond risk parity portfolio against a 60/40 portfolio over time.
Interestingly, the portfolios actually had similar duration for a good deal of the last 40 years. We have shaded in the period from 1970-2005 as the in sample period for risk parity, because that is often the period that risk parity managers seem to show for their backtests. Today is not the first time a significant gap has opened between the two, but it means that the risk parity portfolio today makes sense only if the future is meaningfully more extreme than our Hell scenario and that it is particularly vulnerable to losses should the future turn out to be more similar to history than that. Risk parity has not shown any particular skill in timing its duration exposure as it has generally done worse relative to a 60/40 portfolio when its duration has been particularly long, as we can see in Exhibit 6.
We believe a better solution is to focus on robustness rather than optimality and build a portfolio that will do decently in either scenario. There are no silver bullets available today, but we can work to bridge the gap between the Purgatory and Hell portfolios. One weapon we can make use of that can help is alternatives. I wrote about alternatives in last quarters letter, and as a quick recap, they have the virtue that with a generally shorter duration than traditional assets, their returns are much less sensitive to which scenario plays out. In our Benchmark-Free Allocation Strategy we have approximately 20% in alternatives today, which is at the high end of our historical range for such assets. But unless we were willing to move the portfolio to be predominantly alternatives, there is only so much they can do to bridge the gap. The rest of the work is being done in a less flashy way, through compromise. Our actual portfolios are somewhat more complicated than the simple examples I used here to demonstrate the difference between the optimal portfolios for today, but the gist is the same. If the right portfolio for Purgatory today has about 35% in risky assets and the Hell portfolio has around 60%, our current portfolio has about 50%. This means we are running more risk than is optimal if this is Purgatory and are leaving some return on the table if this is Hell. But the Benchmark-Free portfolio is designed with the goal of having the same expected return in Purgatory as the optimal Purgatory portfolio, only about 0.5% per year worse than the Hell portfolio in the Hell scenario, and has materially less duration, depression risk, and volatility than a traditional portfolio. It may not be strictly optimal for either Purgatory or Hell, but it does allow us to be more philosophical about our uncertainty over which scenario will play out over the next seven years.
Purgatory and Hell, the 30-year view
Over seven years, Hell is a less negative scenario than Purgatory because the capital losses associated with falling valuations are smaller. Over 30 years, however, the positions reverse. Given the longer time horizon, the lower income available from stocks and bonds outweighs the benefit of valuations falling by less. Exhibit 7 shows the 30-year forecasts for the Purgatory and Hell scenarios in both real (inflation adjusted) and nominal terms.
Over seven years, Hell is a less negative scenario than Purgatory because the capital losses associated with falling valuations are smaller. Over 30 years, however, the positions reverse. Given the longer time horizon, the lower income available from stocks and bonds outweighs the benefit of valuations falling by less. Exhibit 7 shows the 30-year forecasts for the Purgatory and Hell scenarios in both real (inflation adjusted) and nominal terms.
The forecasts are worse for every asset in Hell, although for equities (given their long durations) the differences are not particularly large. One point that becomes immediately clear from these forecasts is that earning the 5% real return that endowments and foundations and savers are counting on, or the approximately 7.3% nominal return that defined benefit pension funds are assuming on average, is going to be extremely difficult. While a few institutions have been able to generate the kind of outperformance that it would take to turn these forecasts into an acceptable outcome, it would be impossible for institutions as a whole to do so. Alpha of that kind is predominantly a zero sumgame. It is certainly possible for a well-resourced, talented, and hard-working investment staff to field a group of active managers to outperform their respective asset classes, as evidenced by Yale Universitys endowment returns over the last 30 years. But active managers in aggregate have not, will not, and cannot meaningfully outperform their asset classes in the long run. This means that whether we are in Purgatory or Hell, institutions have a very big problem today.
The nature of that problem is different in Purgatory and Hell, however. In the Purgatory scenario, returns for the next seven years will be inadequate, but after that, the world reverts to a normal pattern and the old rules apply. Exhibit 8 shows the value of a corpus starting at $1 million and with the spending rules of a private foundation in the normal scenario and the Purgatory scenario. This assumes a traditional 60/40 portfolio.
The bad news is that the real value of the corpus falls about 36% over the next seven years as returns are far below the 5% real return required in the long run. But the good news is that it is at least a finite problem. The value of the corpus will fall for a while, but at least at the end of that period it reaches stability again. As my colleague John Thorndike has put it, this means that Purgatory is an investment problem. By this, John means that the basic institutional rules that investors have lived by, with regard to pay-outs, investment policy and the rest, are still valid in the long run. The problem for the next seven years is how to mitigate the pain.
If we truly turn out to be in Hell, the problem is quite different. Exhibit 9 is the same idea as Exhibit 8, but also shows the flight path of Hell over time.
The losses for a traditional 60/40 portfolio are less severe in the Hell scenario for the next seven years, but the trouble does not end with year seven. If we are in Hell, the basic math for foundations simply fails. A 60/40 portfolio has no hope of delivering the 5% real that the foundation needs to pay out over time, and in fact no combination of stocks and bonds will get you to 5% real, because the equilibrium for stocks is only 4.5% real in Hell, and they are the highest-expected-return asset. This means, in Johns terminology, that Hell is an institutional problem. The basic assumption that the foundation could expect to live on in perpetuity becomes false. Translating to the endowment problem, the spending rate that preserves intergenerational fairness has fallen materially. For someone saving for retirement, the required savings rate has just risen markedly, because not only is the pot of retirement savings going to be growing more slowly than expected, but the safe spending rate for a given pool of savings in retirement has just fallen. For a defined benefit pension plan, making reasonable if arbitrary assumptions of a duration of 15.5 and payouts over 30 years, if it was fully funded at a 7.3% expected return, the funded status drops to 80% in Purgatory and 67% in Hell.
Conclusion
Both Purgatory and Hell are bad outcomes for investors, but we believe the basic point is that from todays valuation levels, there are no good outcomes for investors. In the nearer term, maintaining todays high valuation levels implies less pain for investors. But as the time horizon lengthens, the lower cash flows associated with higher valuations bite more and more, so the longer the time horizon, the more mean reversion of valuations is a positive rather than a negative. In the long run, we can hope that valuations fall to historically normal levels, because only if that happens will the institutional business models and savings and investing heuristics that institutions and savers have built still be valid.
Both Purgatory and Hell are bad outcomes for investors, but we believe the basic point is that from todays valuation levels, there are no good outcomes for investors. In the nearer term, maintaining todays high valuation levels implies less pain for investors. But as the time horizon lengthens, the lower cash flows associated with higher valuations bite more and more, so the longer the time horizon, the more mean reversion of valuations is a positive rather than a negative. In the long run, we can hope that valuations fall to historically normal levels, because only if that happens will the institutional business models and savings and investing heuristics that institutions and savers have built still be valid.
But it is difficult to dismiss the possibility that this time actually is different. While all periods in which asset prices move well away from historical norms on valuations have a narrative that explains why the shift is rational and permanent, I would argue that this time is more different than most. Whether we are talking about 1989 in Japan, 2000 in the US, or 2007 globally, what bubbles generally have in common is that the narratives require the suspension of some fundamental rules of capitalism. They require either large permanent gaps between the cost of capital and return on capital, or some group of investors to voluntarily settle for far lower returns than they could get in other assets with similar or less risk. Todays market has an internal consistency that those markets lack. If there has been a permanent drop of discount rates of perhaps 1.5 percentage points, then market prices are generally close enough to fair value that you can explain away the discrepancies as business as usual. That does not mean that the drop in discount rates actually is permanent. Markets are notorious for a lack of imagination about how the future can differ from the present. But the models that purport to explain the natural level of short-term interest rates generally show neither much ability to explain history nor have a lot of theoretical appeal to them. This means we have poor guides for what short-term interest rates should be. Sadly, I believe that includes the model of whatever rates have been on average over the last N years, which is the market historians default model for the future. That makes it hard to dismiss an argument that the rate has changed in a durable way, even if the argument that the change will be durable is a largely speculative one.
Will the future look like the Hell scenario? We hope not. While it is better for the next few years, the longer-run implications are fairly bleak. But hope is not an investment strategy. We believe that today it is irresponsible not to at least contemplate the possibility of Hell in building a portfolio, and likewise it is irresponsible for institutions not to contemplate the possibility of Hell when thinking of their broader decisions. For our part, we work first to try to build the right portfolios for our clients, and today our idea of how to do that is to build a portfolio that is robust to either outcome.
But we also consider it part of our job to help advise our clients beyond the portfolios we run, and the Hell scenario is a possibility that few, if any, investors are truly prepared for. Savers and institutions had a very nasty shock when 2008-09 happened to their portfolios. But in some ways it was easy to deal with. Having a 30-40% hole blown out of your portfolio suddenly causes plenty of soul searching and difficult decisions. But the need for the decisions is so obvious that everyone involved knows something needs to be done, and done fairly quickly. Hell, on the other hand, will creep up slowly. The problem will not be a giant drop in the value of the portfolio, but serial disappointment in returns that doesnt amount to a lot in a single year. Only the accumulation of disappointing years and a slowly shrinking corpus/falling funded status/inadequate accumulation of retirement savings will make clear the depth of the problem.
The trouble is that by the time it is obvious in a 2008-09 sense that something is badly wrong, investors will have wasted a lot of time that could have been used to mitigate the problem if only they had acted. One unfortunate feature of the institutional landscape is that there will be very little praise and no lack of scorn for those who are first to move in the necessary direction to survive Hell. The pension fund CIO who recommends moving expected returns from 7.3% to 4.5% is overwhelmingly more likely to be fired than lauded. But ignoring reality does not change reality. While the circumstances in which Jeremy Grantham wrote Reinvesting When Terrified were quite different than todays daunting prospects for institutions, his advice nevertheless is still completely true. There is only one cure for terminal paralysis: you absolutely must have a battle plan... and stick to it. Even if there is too much uncertainty about which scenario we are facing as investors to know exactly the perfect plan for todays portfolio and for tomorrows institutional dilemmas, now is exactly the time we need to start building our battle plans for the challenges we will be facing in the coming years.
