REM has been paying a dividend yield around 15%.
This level of dividends in inconsistent with the yields of funds with similar portfolios.
My calculation of the next REM dividend suggests a significant decline from recent quarterly dividends.
REM is still a good way to collect high yields, if you are comfortable with interest rate risks. However, be prepared for a reduction in the dividends in coming quarters.
iShares Mortgage Real Estate Capped ETF (NYSEARCA:REM) has a 15.11% yield based on the annual dividends and the recent price of $12.77. REM is an ETF comprised primarily of mREITS. The questions are: how does it pay such high yields? And can it keep doing it? These questions arise, since another ETF, Market Vectors Mortgage REIT Income ETF (NYSEARCA:MORT), contains a very similar basket of mREITs as REM, but has a 9.40% yield based on the annual dividends and the recent price of $25. Both REM and MORT are not leveraged. So use of leverage cannot be the explanation. UBS ETRACS Monthly Pay 2X Leveraged Mortgage REIT ETN (NYSEARCA:MORL) has the same index components as MORT, but uses 2X leverage and thus has about twice the yield of MORT given today's very low short-term interest rates, which results in very low borrowing costs for MORL.
The largest components by weight of REM are the same as those of MORT and MORL. Annaly Capital (NYSE:NLY), American Capital Agency Corp. (NASDAQ:AGNC) and NorthStar Realty Finance Corp. (NYSE:NRF) are the top three components of REM, MORT and MORL, and comprise about a third of the weights of each. The weights of all of the components of REM are included in the table below. REM has more components than the others. REM's 36 components include some higher-yielding mREITs, such as Javelin Mortgage Investment (NYSE:JMI) and Ellington Residential Mortgage REIT (NYSE:EARN), not included in the 24 components of MORT. However, they only marginally account for the much higher dividends paid by REM as compared to MORT on an ongoing basis. I know this because I estimate the dividends that will be paid by various ETFs and ETNs by calculating the number of shares of each component and then multiplying the number of shares by the dividends for each component and then dividing the sum of the dividends by the shares outstanding of the ETF.
The procedure for calculating the dividends is more fully explained in: Projected June Dividend For CEFL Will Bring Yield To 18.1%. It should be noted that in the table below, for the mREITs that pay monthly dividends, such as JMI and Armour Residential REIT (NYSE:ARR), quarterly dividend rates are shown. The quarterly dividend rate is the sum of the three recent monthly dividends.
An example of how the dividends for each component are calculated can be illustrated for NLY. The first step is to take the percentage of total net assets that NLY comprises of REM, which is shown on the table below as 14.74%. Then apply that 14.74% to the total net assets of REM. Thus, REM holds (14.7% x $1,327,193,571 = $195,628,332 worth of NLY). Dividing the value of the NLY position by the $11.54 share price for NLY results in a share count of 16,952,200. Multiplying by the NLY dividend of $0.30 gives a dollar value for NLY's dividend payment to REM of $5,085,660. My calculation indicates that the next quarterly REM dividend, that will be announced soon, should be about $0.30, a sharp reduction from the $0.5067 paid on March 31,2014 and the $0.4767 paid on December 30, 2103. My projection of a REM dividend of about $0.30 would bring the annualized yield based on that dividend to a level similar to that of MORT on an ongoing basis.
The fact that even taking into account the additional components, which my calculation does, REM should yield close to MORT, makes it even more perplexing that REM has recently yielded about 50% more. Examining the most recent Prospectus, Annual Report, Semi Annual Report and Statement of Additional Information for REM eliminates most of the usual suspects that are the typical causes when dividends appear higher than what it looks like they should be. There is no significant return of capital. The numbers all add up, and there does not appear any gains from in-and-out trading or similar strategies. The portfolio turnover was a miniscule 17%.
Adding to the mystery was that the assets held by REM as published in its latest filing, the Semi Annual Report for the six months ended October 31, 2013, showed a portfolio very similar to that shown below. The largest components by weight were NLY and AGNC, just as now. On October 31, 2013 REM had 6.2% in cash, actually BlackRock cash funds yielding 0.14% or less. Now the amount of cash is negligible. The net investment income for the period ended October 31, 2013 was almost all from dividends from mREITs unaffiliated with BlackRock.
The mystery began to unravel when the line item "Ratio of net investment income to average net assets" (annualized) was examined. This is a proxy for yield when dividends are very close to net investment income per share and the market prices of the shares are very close to net asset values, as was the case for REM. For the six-month period, net investment income per share was $0.99, and distributions from net investment income were $0.95 per share. There was no return of capital.
For the six months ended October 31, 2013, the ratio of net investment income to average net assets (annualized) was 15.27%. That is consistent with the recent dividend yields around 15%. However, when the ratio of net investment income to average net assets for previous periods is examined, this appear right in line with MORT. For the year ended April 30, 2013, the ratio of net investment income to average net assets for REM was 10.96%. For the year ended April 30, 2012, the ratio of net investment income to average net assets was 11.44%. For the year ended April 30, 2011, the ratio of net investment income to average net assets was 9.58%. For the year ended April 30, 2010, the ratio of net investment income to average net assets was 9.62%. These are consistent with MORT and similar funds.
Thus, the anomaly of dividend yields higher than what the portfolio suggests is concentrated in the period after April 30, 2013. When the REM dividend of $0.433 was paid on April 30, 2013 with an ex-date of March 25, 2013, on which day REM was at $15.84, the annualized yield based on that dividend of $0.433 was 10.93%. MORT went ex-dividend on April 1, 2013, with a dividend of $0.726; on that day, MORT was at $29.89 and the annualized yield based on that dividend of $0.726 was 9.72% After April 2013, the dividend yields diverged. If my estimate of $0.30 for the upcoming dividend for REM is correct, they will be converging.
The relatively very large (for an unleveraged mREIT fund) dividend yield was due to a number of factors in which the stars just happened to align correctly for REM and those investors who received dividends from it. The timing of the dividends declared and paid by the mREIT components of REM and the market prices of the individual mREITs in response to what has been called the "taper tantrum" in the summer of 2013 combined to allow REM to maintain a high dividend yield as the numerator (the dividends) remained fairly constant while the denominator (the market prices of mREITs) declined.
Additionally, REM benefited by extraordinary dividends that were paid at times that benefited the dividend yield of REM. Western Asset Mortgage Capital (NYSE:WMC) paid an extraordinary $2.35 on January 28, 2014, after paying $0.90 on September 30, 2013. In my calculation, I am using the latest dividend declared by WMC of $0.67 that was paid on April 29, 2014. As with all of the mREITs in the table below, it is possible that different dividends may be declared between now and when REM makes its next payment. However, most mREITs that have already declared their upcoming quarterly dividends, such as CYS Investments (NYSE:CYS) have kept them the same as last quarter. It should be noted that WMC is not a component of MORT and MORL, but is of REM.
Even though it is likely that some investors who are expecting that REM will be able to maintain a 15% yield are likely to be disappointed, I do not think REM is a bad investment. It is still likely to be paying a double-digit yield, although closer to 10% than 15%. REM is somewhat more diversified than MORT and MORL, since it has 36 different issues in its index as compared to 24. It is less risky than holding any of the individual mREITs that comprise the index. While the yield is less than MORL, because it does not have the 2X leverage that MORL employs, that means that the risks associated with the 2X leverage are also absent. These risks include those associated with note structure that allows MORL to employ 2X leverage, which means that MORL is a note that carries some credit and forced redemption risks. See: UBS Leveraged ETNs: Separating Fact From Fiction.
REM does provide a significant degree of diversification relative to individual mREITs. As I explained in: 30% Yielding MORL, MORT And The mREITs: A Real World Application And Test Of Modern Portfolio Theory, a security or a portfolio of securities is more efficient than another asset if it has a higher expected return than the other asset but no more risk, or has the same expected return but less risk.
Portfolios of assets will generally be more efficient than individual assets. Compare investing all of your money in one security that had an expected return of 10% with some level of risk, to a portfolio comprised of 20 securities each with an expected return of 10% with the same level of risk as the single security. The portfolio would provide the exact same expected return of 10%, but with less risk than the individual security. Thus, the portfolio is more efficient than any of the individual assets in the portfolio.
While 90% of REM's portfolio is in agency mREITS, the other 10% is in other types of REITS, such as commercial mortgage financing, industrial and office REITS. The agency mREITs are primarily dependent on the continuation of low interest rates to maintain their dividends. Thus, weaker economic activity is good for the agency mREITS. Some of the agency mREITs do contain some non-agency mortgages, which entails some credit risk. Mortgages that do have credit risk benefit from stronger economic activity and generally suffer from weaker economic conditions. Thus, at least 10% of REM's portfolio will do better in periods of increased economic activity, which provides some diversification for the risks that agency mREITs will suffer when economic activity increases. See: A Depression With Benefits The Macro Case For mREITs.
REM could also be a good investment for those who want higher yields and want to use their own leverage to do so. Buying REM on 50% margin would return a higher or at least comparable yield to buying MORL for those who could borrow at LIBOR plus 0.4% or some similar level. Many retail investors cannot borrow at interest rates low enough to make buying REM on margin a better proposition that buying MORL. However, larger investors with access to low margin rates might do better by buying REM on margin.
Since REM shares can be created or redeemed (in lots of at least 50,000), there is not much risk that there will be a sharp sell-off in REM prompted by the dividend yield returning to the level I am forecasting. Any downward pressure on REM would result in arbitrage in which REM shares were bought to be redeemed if the net asset value was significantly above the market price. Since REM is relatively large, with over $1.3 billion in assets, it is theoretically possible that selling pressure in REM could put downward pressure on the mREITs in the index. However, REM is still small compared to the total capitalization of the mREITs that comprise the index. NLY alone has a capitalization more than $11 billion. This would make the "tail wagging the dog" scenario of REM pulling down the prices of the components less likely.
In summary, REM is still a good way to collect high yields, if you are comfortable with the interest rate risks associated with mREITs. However, be prepared for a reduction in the dividends in coming quarters, as the extraordinary factors that allowed REM to pay more than comparable non-leveraged mREITs funds are no longer present.
REM Components June 11, 2014
ANNALY CAPITAL MANAGEMENT INC.
AMERICAN CAPITAL AGENCY CORP.
NORTHSTAR REALTY FINANCE CORP.
STARWOOD PROPERTY TRUST INC.
TWO HARBORS INVESTMENT CORP.
MFA FINANCIAL INC.
CHIMERA INVESTMENT CORP.
INVESCO MORTGAGE CAPITAL INC.
HATTERAS FINANCIAL CORP.
COLONY FINANCIAL INC.
NEW RESIDENTIAL INVESTMENT REIT CORP.
NEWCASTLE INVESTMENT CORP.
ARMOUR RESIDENTIAL REIT INC.
ALTISOURCE RESIDENTIAL CORP.
CYS INVESTMENTS INC.
REDWOOD TRUST INC.
ISTAR FINANCIAL REIT INC.
PENNYMAC MORTGAGE INVESTMENT TRUST
AMERICAN CAPITAL MORTGAGE INVESTMENT
CAPSTEAD MORTGAGE CORP.
APOLLO COMMERCIAL REAL ESTATE FINANCE
ANWORTH MORTGAGE ASSET CORP.
NEW YORK MORTGAGE TRUST INC.
RESOURCE CAPITAL REIT CORP.
RAIT FINANCIAL TRUST
WESTERN ASSET MORTGAGE CAPITAL CORP.
APOLLO RESIDENTIAL MORTGAGE INC.
AG MORTGAGE INVESTMENT TRUST INC.
DYNEX CAPITAL INC.
ARES COMMERCIAL REAL ESTATE CORP.
ARBOR REALTY TRUST INC.
HANNON ARMSTRONG SUSTAINABLE INFRA CAPITAL
JAVELIN MORTGAGE INVESTMENT CORP.
ELLINGTON RESIDENTIAL MORTGAGE REIT
ZAIS FINANCIAL CORP.
*JMI pays monthly dividends of $0.15, the quarterly rate is shown
*ARR pays monthly dividends of $0.05, the quarterly rate is shown
Disclosure: The author is long MORL, AGNC, CYS, ARR, REM, JMI. The author wrote this article themselves, and it expresses their own opinions. The author is not receiving compensation for it (other than from Seeking Alpha). The author has no business relationship with any company whose stock is mentioned in this article.