BlackRock's Multi-Sector Income Trust is a new fixed income (FI) closed end fund (CEF) with a flexible mandate to invest in a wide range of bond sectors. The vehicle was IPO'd a few months before 2013's bond sell off and there have been fears of unhealthy return of capital (ROC); these two factors have combined so that shares change hands at a considerable discount. The fund's adviser, BlackRock, is one of the larger sponsors of closed end funds, across asset classes and investment objectives - and so far management has been responsible stewards at BIT. The large discount to net asset value (NAV) combined with strong earnings that exceed the current distribution, present one of the better opportunities in the multi-sector FI CEF segment.
The fund has a mandate to invest across global debt markets, and is among the wave of flexible / unconstrained bond mandates that investors are clamoring for. The fund is able to use leverage as well as a spectrum of derivatives. As one can see from recent holdings above there is little it is limited in investing in -- within the bounds of Registered Investment Company laws, and BlackRock's own internal assessments. The portfolio however is significantly different from a simple levered up conventional "core" fixed income fund though. Here is a link to a recent fund card. One can see from the comparison below how sector allocations are different from a sample FI core index, and how that translates into a short term forecast 1 month tracking error difference. Holders of BIT have exposure to many more sectors than in a typical bond index.
The Multi-Sector Income Trust does not have a long track record, but the portfolio management team running the fund has been at the firm (or previously acquired entities) for quite some time, in some cases going back more than 20 years. The quantitative tools to measure effectiveness are less powerful with small data sets, but one can look at a factor loading analysis to understand how NAV returns were generated. [As a side note - for all these kinds of analysis, investors should use NAV returns, not market price returns. NAV is what portfolio management has control over; market prices add two degrees of freedom (discount/premiums, and dilution/buyback) that add noise to the study.] Presented below is a study of the fixed income factor loadings, on a weekly basis, over the last six months.
This suggests that a simple collection of FI beta factors would have only explained 75% of the variation in the fund's NAV movements. In addition, there is both some alpha, suggesting value is being added above simple beta exposure, and a high information ratio, suggesting that when the manager takes risk (with respect to the factor benchmark) it has been rewarded. These numbers are very high, so we'll take them with a grain of salt given the small data set, and perhaps ascribe them to something more prosaic - maybe less alpha, and instead a missing factor from our models. The models do not have a factor which captures the excess returns generated from private label mortgages; this has been a popular alpha strategy in recent years by mortgage oriented hedge funds, and unconstrained bond managers. One can get sense of how management thinks about the market, and their approach to portfolio construction by reviewing these comments they published. After reviewing the numbers, combined with analysis presented further in this note, we'll render a judgment that management is effective, and not responsible for the fund level discount by self-sabotage.
Return of Capital
Although many funds trade at discounts, often warranted, BIT's is larger than peers. Based on discussions with other investors, one recurring theme for the skepticism is the fear of unhealthy return of capital. When a fund makes a distribution that is not 100% covered by earnings, they are required to contemporaneously file a Section 19 notice, disclosing the percentage coming from net investment income. Below is a recent one:
Source: SEC filings
Some FI CEF investors are unwilling to pay for their own money to be returned to them, and rightfully recognize that a high distribution may not all be "yield" if the underlying portfolio cannot cover that. This is a valid concern if a fund is trading a premium to NAV, since by definition, if it is over distributing, it is undergoing a slow moving liquidation, and investors are going to only recover NAV. When a fund is trading below NAV, it may be less of an issue, since it's analogous to a constant self-tender at NAV, yet investors remain skeptical, and because FI CEF holders tend to be distribution driven, fear that the payout will be cut. In addition, from the filings, management also provides an annual disclosure of the ROC.
Source: SEC filings
Superficially, looking at just these two disclosures, one can understand the concern - is the distribution supported?
Earnings & Dividend Support
From the above it becomes clear why investors have stayed on the sidelines. But by digging deeper analysts can develop understanding of the true portfolio earning power. The first thing to remember is the timing of the launch - the vehicle was launched right before a sharp bond sell off -- and was likely to have booked some losses in its early months as it rotated into even higher yielding bonds than what was initially purchased. Those could create the potential to characterize some investment income as ROC, which is being construed as the bad kind by investors. (There is good ROC too, but that's a topic for another day, where MLP & Option CEF enthusiasts can opine with more fervor than me)
Because the fund was launched in 2013, and BlackRock desired to have some synergies from money management it combined reporting with other funds cycles, so it's first semiannual financial report was a stub period. Then, that was combined with the next six months of returns in the full financial annual report, recently released. Because there is a ramp up period when IPO proceeds are held prior to being fully deployed, FI CEFs typically under earn their distribution in their first cycle. Some sponsors resolve this by only starting distributions when a portfolio has been assembled. In any case, to determine the current run rate earnings (looking backward 6 months) for BIT, prospective investors should look at the GAAP net investment income (NII) for the full year, less the stub period.
Source: SEC filings
We can see that the full accounting (partial calendar) year NII was 1.02, and the stub was .16, netting to .86 as portfolio NII over the last six months (from April 30 to Oct 31, 2013). This works out to .1433 per month, versus a payout rate of .1167 - a 20+ % margin of safety. If the earnings rate for the last six months was annualized, it is close to ~9% on NAV, and ~10% on market price - versus a current ~8% market price distribution. That provides protection for the current dividend, and those excess earnings will either accrue into NAV, or (eventually) be swept out each fiscal year as a special dividend if there are no capital loss carry forwards or over distributed net investment income accounting entries to burn off.
Perhaps the last six months was extra profitable, and not representative of future earning power - if so, what other methods can investors use to estimate future earnings for a bond portfolio? A couple of other methods will be used to estimate earnings, since that's one concern driving the discount. First, analysts can take a Yield to Worst / Yield to Maturity (YTW/YTM) approach, by sector allocation, and then adjust for leverage, expenses, and the discount. This is similar to Return on Assets (ROA) analysis for companies undergoing restructuring or transformative change.
The underlying portfolio has an YTW/YTM of 6.20 % and the equity is currently levered up approximately 40 %. We'll assess 150 bps for expenses including the leverage, and then apply the benefits of the 10% discount. We get an approximate earnings power of [(6.2 x 1.4) - 1.5] / .90 ~ 8.00 % -- which is essentially the current yield at market price. This is a safe estimate of what the fund should be able to throw off in sustainable cash flow, assuming rates are stable. This measurement suggests the current dividend can be covered by the existing portfolio.
Because the fund invests in capital securities like preferreds which do not accrue interest in the same accounting fashion that bonds do, the accounting can look lumpy for earnings. (This could be another factor why the fund may occasionally file Sec 19 statements, as the distributions outpace the current cash flow, since preferreds do not use accrual accounting for GAAP income.) The chart above shows the lumpiness over the next 12 months on the existing portfolio; the table below shows the combined amount of interim cash flows over the next 12 months, assuming the portfolio does not change. The 56.6mm coupon/dividend cash flow over the current share count of 38.4mm shares covers the funds distribution rate annualized. To be fair, that does not include a charge for fund expenses, and leverage costs, so some might object to that metric of coverage, but that is also ignoring the sizable amount of non-agency residential mortgages (25+ %) the fund holds, which are trading below par, and, as the housing market recovers, periodically prepay at par as loans get refinanced or home owners move. Those events create a small windfall that goes into the principal cash flow category which is not being included in this coverage measure, even though the fund is unlikely to have paid par for their non agency RMBS holdings.
Based on these three measures (backwards looking stub period adjusted earnings, forward looking YTW sector allocation, and forecast cash flow distributions) there is a reasonable basis to believe the existing distribution is indeed being earned, and is sustainable, despite the periodic filing of section 19 statements. The Return of Capital characterization of some of the distributions seems unlikely to be an economic (bad) ROC, and more likely to be a function of market volatility exploited around the funds launch.
Because the fund holds a range of preferreds, which tend to generate distributions that are, from a tax perspective, qualified dividends, the after tax equivalent yield for holders in a taxable account is slightly better than the stated yield. After the fund normalizes and ROC tax treatment is over, it's reasonable to think that the pfd allocation (plus some wiggle room) will be a reasonable estimate for what the QDI treatment will be on an ongoing basis. Most multi-sector funds do not have allocations to capital securities like preferreds, TRuPS etc. Below is the 2013 QDI estimate.
Source: SEC filings
Risks from Interest Rate & Currency Movements
Bond investors were given a vivid example during the early summer of 2013 of how bonds may not be as low risk as popular perception might be, during a rising rate environment. One measure of interest rate risk is called duration, which is a often expressed as a measure of how much a portfolio will move for a 1 or 100 bps move in interest rates. Typically this is shown for a parallel shift in the yield curve, which is a situation where short, intermediate, and long rates all move the same amount. Although that's unrealistic, it's a simple way to quickly get a sense of the risks a portfolio is carrying versus rates movements. Between May 1 and July of 2013, US 10 year rates, effectively moved 100 bps, providing a shock to measure portfolio sensitivity. The NAV of a CEF fund will be sensitive to both the duration of its holdings, and the leverage that has been applied. Below is a chart showing the fund's NAV movement, against ETF proxies for US 10Y bonds.
Fund management was probably somewhat surprised by the movement, as they had been suggesting their duration was in the 2-4 range, when effectively, the fund ended up performing with an empirical duration closer to 6, which is not that much different than a ten year note. During the taper tantrum, many spread sectors experienced not just a shift in the underlying rates curve, but also a re-pricing of the spread required over the risk free curve. This was particularly sudden in the mortgage sector, where the fund had large allocations. What's interesting is now measuring what happened next. In late November there was another sharp rate shock when US10Y moved ~15 bps over the course of a day or two. Examining NAV movement below, again compared to benchmark UST 10Y ETFs, shows much less sensitivity, scaled to be approximately a duration of 2 now. So perhaps hedges were put in place, the portfolio was tweaked, or the spreads that were most sensitive to rates had already experienced all the widening they were going to. In any case, it suggests that the vehicle is now more protected against the risks of rising interest rates, if the new NAV movement under the rate shock is representative of future behavior.
In addition to seeing how NAV is affected under historical events, one can also conduct stress tests using risk management software. These tests are on the portfolio, so the results would need to be scaled up by the amount of fund level leverage, which can be range from 25% to 40% depending on positioning. This analysis does not include the effects of off balance sheet derivatives, or funding structures like preferred issued by the fund, or repo finance. Below is an estimate of how the portfolio would move in total if US rates moved, as well as a decomposition of that, showing which allocations would be most impacted. Note that some of the preferreds are grouped into the corporate category which combines both bonds and junior securities.
Source: Bloomberg (both images)
There is another rate scenario that is worth investigating - the funding of the leveraged part of the portfolio. From SEC filings analysts can observe that the fund carries 550mm of borrowings, at a cost of 40 bps. If short rates became unmoored and floated higher -- although forward curves give no indication of this for at least two more years -- what might the impact be? Using a 100 bps shock, one can estimate 550mm x 100 bps = 5.5mm. Dividing that by 38.4mm shares equals a 14c hit to earnings, a shade over a cent per monthly distribution, and equal to about a 10% reduction in the dividend. This assumes that the rest of the curve does NOT move, which is not a realistic assumption. (Turkey in fact recently has seen an inversion in their yield curve, where short rates are now higher than long rates, and all rates went up but in different amounts).
Source: SEC filings
There is another sensitivity to investigate - emerging markets (EM) exposure. Because many multi-sector bond funds have exposure to global, especially high yielding EM paper, it would be helpful to see what exposure the overall portfolio has. Note that some sectors, even if not directly issued by EM sovereigns/corporates, may end up having sensitivity to that still, as risk appetites change and liquidity dries up. Below is a chart showing BIT NAV movements over a recent EM strong "shock week" during January 2014. Comparisons were run versus proxies for both local and hard currency EM bonds; from the chart it seems that the primarily domestic oriented BIT was not overly affected.
The Multi-Sector Income Trust appears to have controlled exposure to short and long rates currently, and does not appear to have exposure to EM risks. These risks are less than most peers, and although not presented here, realized NAV volatility is also less than other comparable funds. These attributes all make sense as the advisor, BlackRock, is well known for their forte in risk management.
In conclusion, the BlackRock Multi-Sector Income Trust offers a secure dividend, covered by the existing portfolio. Valuation is attractive versus other flexible multi-sector closed end bond funds given the 10+ % discount shares change hands at, and an 8% current distribution. The discount is unwarranted because risks are controlled, management is competent, expenses are fair, and the fund is large with good share liquidity. The sec 19 filings have created excessive concerns, and when combined with the recent launch right before the taper tantrum, led to BIT becoming a busted IPO. There are many other recently launched funds from sponsors with similar mandates, but BIT has a deeper discount, and an equally high amount of earnings per unit of NAV volatility. Because of the slice of QDI, this may be suitable for taxable accounts as well.