用于长期投资的杠杆ETF

Market history has shown that an allocation to bonds can diversify and lower the risk of an equity portfolio.

However, since bond returns have historically lagged equities, an allocation of investor capital to bonds has also lowered overall portfolio returns.

Leveraged ETFs can allow investors to maintain their desired equity allocation while freeing up capital for bonds.

We investigate historic returns on leveraged equity and bond portfolios and find that 100% SPY + bond portfolios have had higher total returns with lower beta and drawdowns.

While the last decade has favored a leveraged equity and bond allocation, we think that forward returns are unlikely to match those of the recent past.

The S&P 500 index is probably the most frequently used market benchmark in the world and many investors, particularly those who operate within equities, compare their investment results to this index. And as many investors have found out, outperforming this index in any sustainable fashion is difficult.

One way to avoid underperforming the S&P 500 is to allocate one's full capital to an ETF like SPY that tracks the index which will try to get as close to the performance of the index as is operationally feasible for a small fee. While this investment approach will avoid the disappointment of underperformance, it suffers from two drawbacks. First, it leaves no capital for any other investment opportunities and secondly, it generates a relatively paltry income stream.

A way to address these two drawbacks while maintaining equity exposure to the S&P 500 is to utilize leveraged ETFs that track the same index while putting the rest of the capital into fixed-income ETFs. This type of investment portfolio typically not only achieves the full desired equity exposure but also holds the promise of 1) outperformance and alpha generation, 2) diversification and a decrease in the beta of the portfolio and 3) increase of portfolio income.

In this article we review this investment strategy of pairing leveraged equity with fixed-income ETFs, how this strategy has fared over the past decade and our views about its future performance.

Our key takeaways are that first, high-quality, long-duration fixed-income investments have proven to be good complements to stocks. Secondly, pairing bond ETFs with leveraged equity funds has historically delivered higher absolute and risk-adjusted returns. And thirdly, at this stage in the investment cycle leveraged equity and fixed-income portfolios are not as attractive as they have been previously, owing to higher leverage costs as well as higher valuations in both stocks and bonds.

We are far from the first to comment on combining equities and fixed-income allocations without explicit leverage. A number of other commentators have written on the topic in some depth. PIMCO has even built a suite of investment products, including the closed-end fund (PGP) around the concept.

The idea of combining stocks and bonds to achieve diversification is probably as old as financial markets. What is new is the investment vehicle that allows investors to do so without explicit leverage. As we suggest above, the difficulty with allocating precious capital to bonds is that doing so comes at the expense of long-term total returns since the compensation for the equity risk premium has historically been higher than compensation for the duration and credit risk premia. This makes investment funds that allow investors to allocate to bonds without sacrificing their target equity exposure new and exciting.

Leveraged ETFs generally provoke a lot of concern on the part of the market commentariat. Some of the concerns are valid but many are not. The typical comment one hears is that these products are only for short-term traders or that they will necessarily underperform either the unleveraged alternative or their leverage factors. These views are incorrect. Much of the criticism is focused on volatility decay which Dane has already covered well here. Suffice it to say that leveraged funds have performed better than most of their critics have suggested.

It is clear that there should be greater care taken in allocating to leveraged ETFs, however, in our view the structure itself can be appropriate for long-term investors with some caveats.

In our view,

the following considerations should apply to leveraged ETF investments

:

Apart from these considerations that focus on the downsides of leveraged products, however, it's worth keeping in mind that they also have advantages. The no-recourse nature of the products can make them more appealing to trading on margin. A related advantage is that these products can deleverage during a sharp bear market such as the one we saw during the crisis. For example, the SPY drawdown in that period was -51% while the SSO drawdown was "just" -81% - well below its 2x leverage. While a return of -81% is nothing to write home about, it's of course much preferred to 102%.

We use the funds below in our portfolio backtests.

Leveraged Equity Funds

These are popular daily reset ETFs that track the S&P 500 index. There are many other different leveraged S&P 500 products that we do not explore here. However, we would specifically point to the monthly reset Direxion mutual funds such as (DXSLX) which may have lower volatility decay. (SPLX) unfortunately ceased trading and (SFLAF) does not seem to be very popular.

Bond Funds

We test the following bond funds in the portfolio.

These are the basic statistics for this fund population:

We now run the total return analysis for portfolios of leveraged equity and bond funds. We start the analysis in 2010 because that is the date when all these ETFs were trading. The legend in the chart is sorted by total return.

Each portfolio targets the same equity exposure as a fully-invested SPY portfolio. In other words the 2x leveraged ETF SSO gets a 50% allocation (with 50% going to a bond fund) and a 3x leveraged ETF UPRO gets a 33% allocation (with 67% going to a bond fund). The portfolios are effectively daily rebalanced though, of course, in practice rebalancing would be less frequent. The table below shows various portfolio statistics - the portfolio that holds just SPY is marked in red.

What do these results tell us?

So, these results are fairly promising - we seem to achieve our goals of 1) outperforming SPY, 2) doing so with a lower beta, lower volatility and drawdowns and 3) at a higher income level than SPY.

The obvious question here is -

is all of this due to the strong rally in rates since the crisis

? The answer here is yes and no. We rerun the total returns starting from September-2011 to May-2019 when 20-year Treasury yields were both about 2.5%. While the level of outperformance decreases, the same number of funds outperform SPY as before.

As everyone in the markets knows, past performance is no guarantee of future returns. So, how do we think about these strategies from where we are standing now?

The first and perhaps most obvious consideration is valuations

. If we are going to utilize leverage to allocate to various assets, the valuations of these assets matter a great deal. On this front, the news is not great. In the chart below we plot the two basic valuation metrics for stocks and bonds: the S&P 500 P/E and the 10-year Treasury term premium since 1990 (the date since which we have term premium data) with the red dot showing the most recent values. We can see that the term premium is close to 30-year lows while the P/E, while elevated, is closer to the middle of the pack.

If we take a longer perspective on the P/E, we can see that it is actually quite elevated now and arguably the middle-of-the-pack reading since 1990 is skewed by the Tech bubble.

The second consideration for the strategy is the relationship between bonds and stocks.

The strategies rely on the ability of bonds to diversify stocks in order to lower the overall beta and drawdowns of the portfolios. In the chart below we plot the rolling beta of TLT to SPY. We can see that the beta has risen over the last decade and is at a fairly elevated level right now although it is still in the negative territory.

We think there are a number of risks to watch here that may limit the ability of bonds to diversify stocks. First, the tight yield valuations limit how much rates can drop further in case of a sharp equity sell-off. And secondly, the heavy reliance of the current market on the Fed means that a monetary policy mistake may be met by both rising yields and falling stocks.

The third consideration has to do with return drivers of the leveraged ETF portfolios.

At its most basic, the relative performance of these portfolios versus SPY has to do with two things: 1) the relative performance of the leveraged equity fund to SPY and 2) the additional income generated by the bond allocation.

Let's see how the first component has fared over the last decade or so. The blue line in the chart below is the annual return differential between a 50% SSO portfolio and SPY, the orange line is the 1y moving average and the red line is the cumulative average. What the red line tells us is that in the post-crisis period SSO has underperformed SPY by about 1% per annum on a leverage-adjusted basis. In other words, if the SPY delivered a return of about 13% per annum, then SSO delivered a return of about 2x12% per annum. What this means is that we need the bond portion of the portfolio to add above 1% per annum in order for the broader portfolio to outperform SPY, all else equal.

What we do in the chart below is add the 1-year moving average orange line from above to the TLT trailing-twelve month yield. We can see that this line has just gone through zero meaning that the TLT yield has not been sufficient to make up for the SSO leverage-adjusted underperformance relative to SPY. Why has this happened? We think the primary reason has to do with the flat yield curve and the fact that long-term yields are now insufficient to make up for the costs of leverage and higher management fees (0.90% for SSO vs. 0.09% for SPY).

Now, this is not an argument to dump the strategy. As we discussed above the value of this strategy is twofold - to derisk the equity portfolio and generate yield. Even if the strategy is not able to generate income in excess of additional fees, the lower volatility, drawdowns and beta may still make it attractive.

In this section we take a closer look at the SSO + TLT strategy, that is, the portfolio that is allocated half to SSO and half to TLT. Both SSO and TLT traded before the crisis and this version of the strategy strikes us as a good mix between leverage and diversification. The other reason we like this combination is that first, the risk parity weight of TLT in a portfolio with SPY is 55% or close to 50% we use here. And secondly, at 11% TLT has one of the lowest percentage of months where both it and SPY fell. The metrics for other funds are shown in the first table above.

Let's first have a look at how the portfolio fared during the crisis. The performance is pretty good - some of the outperformance versus SPY comes from SSO deleveraging as well as a rally in bonds.

The chart below breaks this down by annual returns. Interestingly, the strategy underperformed in 2018 - this was because yields actually finished higher that year.

The chart below breaks down total returns of the portfolio by asset - showing, in particular, how TLT offset the drop in stocks in 2008.

Finally, we show a 1y rolling drawdown chart of both portfolios. The chart shows how the SSO+TLT portfolio had lower drawdowns in the first part of the decade, however, this has been less the case more recently.

Are there funds that follow these strategies already? One manager that is fond of these strategies is PIMCO which has a whole suite of funds that allocate to equity indices with a bond overlay. We include a number of these funds against several strategy portfolios below.

PSLDX - an institutional-class StocksPLUS Long Duration Fund mutual fund has delivered a 20% return since 2010 by investing in S&P 500 futures alongside a long-duration bond portfolio. The fund matches up pretty well against the UPRO + ZROZ strategy which has a slightly lower sharpe ratio.

PSPAX - a class-A StocksPLUS mutual fund which allocates to mortgages and shorter-term credit doesn't appear to deliver much excess performance to SPY since 2010 or since the inception of SPY.

PGP - the Global StocksPLUS & Income closed-end fund NAV matches up fairly well against the SSO+TLT strategy although the 21% PGP price premium is not attractive at present.

So, what do we think here?

We think the leveraged ETF strategy has shown to be fairly resilient through the crisis and has performed very well over the last decade. Current valuation levels of both stocks and bonds as well as the high cost of leverage relative to long-term yields make the strategy less attractive currently. Positive covariance risks such as monetary policy mistakes that can drive both assets lower are additional risks to the strategy. That said, we think allocating a portion of the portfolio to this strategy may make sense by adding systematically on drawdowns. We prefer diversifying across both stock and bond funds while overweighting longer-term investment-grade funds like VCLT or medium-term treasury funds like IEF. We would also avoid using leveraged bond funds at present given the high cost of leverage relative to long-term yields.

I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.