Since May of this year, mortgage real estate investment trusts [mREITs] have been deeply and negatively impacted by talks of tapering of QE3 by the Federal Reserve. Each time stock prices experience a steep decline, a contrarian ponders whether people over-reacted and whether this constitutes a buying opportunity. This article aims at showing that it usually pays off to be a contrarian, and to do so I apply well known event-study methods to the mREIT case to test average market reactions to past sell-offs.
The recent steep drop in mREIT share prices was justified by the fact that long-term interest rates rose suddenly and violently, and are likely to continue their ascent (hopefully in a more orderly fashion). This resulted in a loss in book value for most mREITs. Ever the optimist, I believe long-term interest rates should rise more slowly from now on, while short-term rates should stay close to zero for the foreseeable future, as the Federal Reserve pursues its ZIRP (zero interest-rate policy). The recent increase in duration and notional amount of interest-rate swaps used by many mREITs should protect book value in the event of a slow rates rise, while the yield curve steepening should result in an increase in net interest spread and income in the coming quarters. The higher interest income should more than balance the increased cost of hedging. Eventually, dividends should rise. You could argue that this kind of argument is akin to gambling: it is simply a bet that the future direction and speed of interest-rate changes can be known. This does not constitute a sound argument for investing, and therefore any mREIT allocation based on such an argument should remain relatively small in a portfolio.
Still, I did add some mREIT shares after prices plummeted. While there is a sound basis for the price drop, I believe the sell-off was overdone. On a general basis, and I am not specifically referring to recent events, I think most investors over-react when it comes to mREITs and have a tendency to sell first and ask questions later at the slightest indication of trouble. In short, I think most sell-offs of mREIT shares are a buying opportunity. The goal of this article is not to invite people to indiscriminately buy mREITs' stocks each time there is such a sell-off, but it is to offer some perspective and remind investors that historically people do tend to over react, and that being a contrarian when that happens usually pays off. MREITs seem to experience sell-offs on a regular basis these days, and I thought it would be interesting to determine the average behavior of the stock return after such an event.
An Event Study
To this end, I performed an event-study analysis following the now standard methodology of MacKinlay (1997) which is well known by academics and institutional investors but maybe less so by retail investors. Event studies have been around for a while (e.g., Fama et al., 1969). The analysis was carried out with the QuantSoftware ToolKit (QSTK) package of the Georgia Tech Research Corporation. The event is specifically defined as a daily drop in stock prices by 4% or more, while the overall market (as defined by the SPY ETF) ends the day stable or higher. This 4% drop was selected somewhat arbitrarily, to result in a large enough event count (but the conclusions of this article remain valid for bigger drops). The adjusted prices at market close provided by Yahoo were used: therefore the price drops resulting from a stock going ex-dividend or from splits are excluded from the analysis. To collect as many events as possible, all of the stock symbols in the REM ETF were used, as long as the market capitalization of the corresponding mREIT is larger than $1 billion (smaller capitalizations experience more price volatility). Consequently, the following symbols were included in the analysis, by decreasing order of market capitalization: NLY, AGNC, STWD, TWO, CIM, MFA, IVR, NRF, HTS, NRZ, RWT, ARR, PMT, CYS, CLNY, CMO, and MTGE. It is a mix of residential, commercial, agency, non-agency mREITs, etc... I also performed an event study on a single company, Annaly Capital Management, the bellwether of the industry.
Event studies aim at determining the impact of a given event on the returns of a stock. Here the event is a sudden price drop resulting, e.g., from quarterly results worse than expected by the analysts, from the announcement of a dividend cut, from a downgrade, from an announcement of a capital raise... It is important to note that this is a mixed bag of negative events: some events (dividend cuts) are more negative than others (capital raise that can be dilutive). Abnormal returns result from this event. Here, following standard practices, I use the market model of normal returns, i.e. I assume that the normal return of the reference portfolio of mREITs is the market return (from SPY) times the ubiquitous beta parameter (set to one for simplicity). The time frame considered for the study is from January 1, 1997, to September 14, 2013.
Result For The mREIT Sector And NLY
This description results in 156 events. The market-relative cumulative returns (abnormal returns) are shown on the plot below, in a 120-day window centered on the event (in other words, time t=0 corresponds to the day of the event). It should come as no surprise to most readers that this plot validates the hypothesis of over-reaction. The day of the event, the share price of mREITs in the study drops, on average, by about 6%, before recovering most of the loss in 50 days or so. On the plot, the days are trading days, not calendar days. On average, the stock price starts recovering within a few days of the event. Moreover, an investor who purchases some shares of mREITs after the sudden drop in price is likely to make a 5% or so capital gain within 50 trading days or less. On the figure the error bars are not shown, but they are quite large (due to the small event count): therefore the strategy of buying mREIT shares after a steep drop will not consistently result in a capital gain, and it is doubtful that such a strategy should be implemented in the first place. However, it does demonstrate that investors have a tendency to over-react, and that more often than not mREIT stocks recover quite nicely and rapidly. This result is by no means surprising, as it roughly matches figures 2a and 2b of MacKinlay (1997) for the overall market (but using a different event definition).
If we now look at a specific mREIT, Annaly Capital Management, it is more difficult to draw firm conclusions due to a smaller event count (only 16 for a 4% price drop). Also we need to define a less restrictive event: e.g., a 2.5% drop in the daily share price, which results in a higher count (55 events). Annaly is the largest mREIT by market capitalization and has been operated since October 1997. The figure below shows that, like other mREITs, the stock price rises on average relatively quickly after a 2.5% or more drop. In fact, after only 40 trading days the stock price is back to its pre-event level. This should be food for thought to anyone holding Annaly in his/her portfolio, even though the usual disclaimer applies (past results are no guarantee of future results, etc...) and current market conditions are very unusual as they result from the expectation of rising interest rates while the long-term rates were mostly decreasing over the time frame considered here.
Comparison With The Overall Market And Another Stock Sector
The same event study, for a 4% drop, applied to all of the stocks in the S&P 500 (with the year 2012 composition) and over the same time interval (January 1, 1997 to September 14, 2013) produces the following figure based on 10726 events:
A word of caution: because the stocks included in the study are those in the S&P 500 from 2012, there is an obvious survivor bias: stocks that may have had daily drops by 4% or more in the past but went bankrupt or left the S&P 500 are not included.
As is known from the academic literature, investors often anticipate bad news responsible for the event studied, even though that seems to be more the case for the whole S&P 500 than for mREITs. Therefore, most stock prices are already declining prior to the event. However, the sell-off occurring the day of the event is overdone, as investors tend to panic, and stock returns quickly climb back. Overall, a contrarian investor buying some shares the day the price drops by 4% or more while the rest of the market is stable or increases is likely to score a 7% capital gain or so within 60 trading days of the drop. Again, it's an average behavior and it does not mean that all stocks behave this way all the time, or that buying each time there is a drop is a sound strategy. It just demonstrates that people over-react more often than not, and that as an investor you should remember this fact next time mREIT prices experience their now-regular sell-off (even though it is always difficult to ignore the noise of articles and blogs describing an impending doom in the mortgage-backed securities market).
Is the event study shown here useful since mREITs and the broad market appear to behave roughly the same way? Some specific stocks do not, on average, climb back after a sudden drop. Therefore mREITs do not behave like all the other stocks. Indeed, the same event study for different stock categories does not always show a strong rising trend. As an example, I looked at the stocks in the XLK SPDR technology ETF (in existence since December 1998). In the study, only 27 stocks from this ETF are included (e.g., AAPL, MSFT, GOOG, IBM, T, CSCO, INTC, VZ, EBAY, etc...). This is a mixed bag of very different companies, and it covers a wide range of the technology sector. The result of this event study is different from what was obtained for the mREIT sector. The 1997-2003 time frame initially selected for the components of XLK includes the dot-com bubble burst of early 2000, so the result is necessarily biased. Also, I later restricted the study to dates after January 1, 2003, and the result is less lopsided but still quite telling:
Based on 244 events, it appears that for technology stocks, on average, it is better not to buy after a sudden price drop because the potential capital gain is lower than in the mREIT case and the stock only recovers a small part of the loss. In other words, the initial sell-off is less overdone. Where does this difference come from (assuming it is a significant difference, which is not clear due to the small event counts)? It might be that, overall, investors understand technology stocks better than mREITs, which are quite complicated stocks to invest in as they deal with rather obscure financial instruments and employ convoluted hedging strategies. Therefore, when bad news occur, investors may be better at estimating the new fair value of a technology stock while they tend to exaggerate the negative impact of the news on mREITs' shares. Moreover, a lot of mREIT "bad" news is actually a basic capital raise: investors assume it will be dilutive and sell some of their shares.
The goal of this small event study is to remind Seeking Alpha readers of a well-known truth: retail and institutional investors over react to bad news. When dividend cuts are announced, when quarterly results fall short of analyst consensus, when an mREIT is downgraded, when the Fed mentions a change in interest rates, etc, the ensuing sell-off in mREIT shares appears to constitute, more often than not, a buying opportunity. Nothing new here, but a little bit of perspective. Moreover, mREITs often announce capital raises, which mechanically result in a significant price drop but are typically a very good time to buy. Of course, not all sell-offs are overdone and/or unjustified, and not all mREITs behave the same as the category encompasses quite different companies (some mREITs are currently much less risky than others as they have less sensitivity to rising interest rates). Therefore a strategy based on systematically and indiscriminately purchasing mREIT shares after a 4% or more sell-off is most likely not sound. Still, on average and over the past 16 years, it appears that being a contrarian and adding shares to your mREIT portfolio after a sell-off led to quick and decent capital gains. The conclusion though, is that investors currently holding some mREITs in their portfolio should take solace in the fact that there have been numerous sell-offs in the past, that overall they were nice buying opportunities, and that the stock prices partly recovered. For sure, the recent sell off was bigger than just a 4% drop, and the current environment of rising long-term rates is different from what we experienced this past decade and creates huge challenges to the sector. Therefore the results of this study might not readily apply to current conditions. Still, I believe that keeping your calm and adding mREIT shares to your portfolio on dips is probably a smart move on the long term, especially for less risky mREITs (like CMO, TWO, etc...) or well managed ones (MTGE...). MREITs do have a place in a portfolio, as long as the positions are kept relatively small (this remains a risky sector after all).