We were recently asked to look at and give our opinion on mortgage investment corporations, or MICs. Given the ultra-low interest rate environment, high yielding MICs have become a popular product among retail investors. What we found behind these high yield products was concerning to us, particularly given the potential for a generalist investor to significantly underestimate the actual credit risk of certain MICs.
Disclosure: At the time of writing, no HCP Fund, account or staff member had exposure – long or short – to any MIC investments.
Our background in banking reminded us of certain lessons from the recent U.S. credit cycle that might help advisors and investors better identify those MICs with lower credit risk and stable yields (versus those which are taking considerably more risk). In our opinion, understanding these lessons is extremely important, if not urgent.
Although our firm does not offer any yield-oriented or mortgage funds, we do specialize in financial services, and a large swath of our investment universe hold mortgages on their balance sheets. Hence the issues relating to mortgages are, in fact, very familiar to us.
It should go without saying: higher loan yields have higher credit risk. If you believe a MIC investment is lower risk because (a) its portfolio is protected by a “low” loan-to-value (LTV) ratio, (b) credit losses experienced by the MIC have been essentially zero over its life (i.e., it has “good management”), or (c) it is “residential”, then please continue reading.
Two Ways Investors Might Underestimate the Risk Profile of a MIC
There are a lot of MICs, and we did not attempt to review all or even most of them. However, for the several MICs we did review, we observed two areas where a generalist investor may – incorrectly – surmise that a MIC investment has low credit risk, when in fact the opposite could be true. These potential misunderstandings are more significant in a market where property values are declining.
First, we suspect many investors may believe MIC portfolios are dominated by lower risk owner-occupied residential mortgages, owing to legislated minimum requirements on “residential” content. However, for some MICs we found, the “residential” content was primarily, if not exclusively, higher risk residential construction/development loans, including the riskiest category of all – undeveloped land.
Second, we believe many investors overestimate the protection (of the investment and its very high yield) offered by a “low” loan-to-value ratio. As we will explain, it is not sufficient for investors to accept the safety of a MIC’s yield based on this metric, which in many instances incorporates a significant number of subjective assumptions (more on this later).
Unfortunately, many market comments we reviewed arguably amplified these two potential areas of misunderstanding (amazingly, one even described MICs as “relatively risk-free”). For some MICs we reviewed, the risk profile of the underlying mortgage portfolios skewed overwhelmingly to higher credit risk, be it through significant exposure to the most risky loan categories, excessive concentration in individual names/geographies, and/or subordinated collateral positions. This was particularly true for those MICs with outsized exposure to construction and land development. In our view, the significant credit risk of certain MICs has been obscured by a robust real estate market supported by large increases in home prices/property values.
Relating U.S. Home Price Declines to Canada
With property values in Canada having risen substantially in the past several years, many market commentators believe that home prices are set to decline. It is obvious that the federal government is very concerned about a hard landing in home prices, and is – through policy actions and words – trying to engineer a soft landing.
With a possible decline in home prices at hand in Canada, it makes sense to look to the U.S.’s recent experience for lessons learned. Unfortunately, many try to extrapolate the experience of the U.S. subprime lenders to that of MICs. However, in our view, subprime lending is not a relevant comparison for many reasons, the most important being that most MICs have limited exposure to owner-occupied residential portfolios. Another important reason is that the U.S. housing market was overwhelmed by a catastrophic decline in underwriting standards and fraud on the part of both borrowers (in misstating their income, assets), and lenders (primarily five or so largest thrifts that originated mortgages while knowingly violating their representations and warranties to buyers of these mortgages via securitization).
In our opinion, a better comparable is the experience of the U.S. construction-heavy banks, as their loan portfolios much more closely mirror many of the MICs we reviewed. These banks, with their excessive exposure to this highly economically sensitive loan category, generated high returns/yields with minimal losses until home prices began to decline. Once property values declined, they suffered meaningful losses, with hundreds of these banks failing.
To be clear, we are NOT suggesting any MICs will fail (or anything close) since: (a) banks have substantial financial leverage (while MICs have very little), (b) fraud is not pervasive in the Canadian market, and (c) the U.S. suffered a severe recession coincident to the decline in home prices, which no doubt exacerbated losses to development projects (and an economic downturn does not seem likely in Canada).
However, we do believe there are very important lessons MIC investors can take from the experience of these banks, even if the order of magnitude will obviously be less severe. This is especially true since many of the reasons MICs and their supporters cite to explain their risk profile sound a lot like those that the U.S. construction banks used before they suffered substantial losses.
Four Lessons MIC Investors Can Take from U.S. Construction-Heavy Banks
It is obvious that MICs vary by their portfolios. Some have exposures dominated by mature multi-family and/or loans against student housing. For these MICs, their yield is supported by diversified cash flows, with limited economic sensitivity. However, others have significant exposure to construction and land development loans, for which cash flows are far less regular or certain.
Therefore, without commenting on the risk-reward attributes of any particular mortgage investment corporation, we would offer four lessons from the experience of the U.S. construction-heavy banks for investors to consider when reviewing MICs with significant exposure to construction, land development, and raw land. With Canadian property values set to possibly decline materially in certain urban markets, if not nationally, we believe understanding the following lessons is very important – if not urgent.
Lesson #1: Not All Real Estate Loans or Portfolios are Created Equal
There are a lot of different real estate loan categories with differing levels of risk, owing to differing cash flow characteristics and collateral. MIC investors can however follow some simple rules of thumb. Owner-occupied (i.e., primary residences) loans are usually less risky than investment or income producing properties (i.e., a condo being rented). Real estate in mature urban areas (i.e., the city) may hold its value better than new developments in the suburbs. For obvious reasons, first mortgages are less risky than second or third mortgages (who needs a third?). Low loan-to-value loans are – generally speaking – less risky than higher loan-to-value loans.
As real estate loans themselves differ, so do portfolios of such loans. Concentration is very important. In general, the higher the exposure to a particular loan category, client (or client group), or specific geography, the higher the risk of losses. Thus, the more diversified the underlying cash flow stream of the mortgage portfolio, the better. In the last cycle, banks most concentrated in states like Florida, Arizona, Nevada, and southern California suffered greatly. As for MICs, those with the greatest exposures to the most speculative markets, or most economically sensitive loan categories, will potentially fare the worst.
Bottom Line: Many MICs we reviewed had loan portfolios which skewed overwhelmingly to concentrated portfolios with higher credit risk. One MIC had ~25% of its portfolio in just 4 mortgages (all seconds) and 40% of its total portfolio in second mortgages. Another had ~40% of its portfolio in land and construction loans.
Lesson #2: Construction/Land/Development Loans Have – by Far – the Highest Credit Risk
The most important lesson learned from the last U.S. credit cycle was that one loan category – construction/land/development – stands out as having a dramatically higher risk profile than all other mortgages/real estate loans (even subprime). And various sub-categories within construction loans also have different levels of risk: vertical construction (high), acquisition and development (higher), raw land (highest).
Unlike other real estate loans, construction loans do not generate consistent cash flows (such as principal repayment/interest) from a third party tenant/occupant of a property. Instead, these loans are often repaid with a balloon payment once the project/home is completed (i.e., the homes are sold, and/or the commercial property has been built and sold). In the meantime, the principal and interest payments to the lender are funded out of the proceeds of the initial loan.
As a result, the collectability of these loans can potentially be very economically sensitive. In a robust real estate market, with lots of volume and rising real estate values (which Canada has had for decades), developments get built and sold, and lenders are repaid without issue.
However, if property values begin to decline, a number of problems may emerge: sales may not take place, pre-sales may walk away (depending on the size of their deposit), and/or the property may be sold at a distressed price. In extreme instances, the developer could go bankrupt because delayed sales cause it to run out of cash before it makes enough sales (homes) or sale (commercial) to repay the loan.
Bottom Line: Many MICs we reviewed had excessive exposure to construction, land development, and undeveloped land (the riskiest of all). It is this significant exposure to these high risk loans that supports the payment of the high(er) yields of many of these income vehicles.
Lesson #3: Absence of Historical Losses is Not Necessarily Evidence of Low Credit Risk
One of the most important lessons learned from the last U.S. credit cycle is not to take any comfort from the lack of historical losses in a portfolio. A market characterized by rising property values and robust new home sales can obscure very high credit risk. As the chart highlights, losses for construction loans held by U.S. banks were essentially zero for the years preceding the downturn. However, at almost exactly the same quarter that home prices began to decline, losses in construction and development projects began to emerge.
And as property values continued to decline, and the recession deepened, losses rose substantially, peaking at 8.4% annualized in Q4-2009. Note that the loss rates depicted in the chart are a national average, and areas with the highest home price increases and most speculative lending activities (and fraud) experienced even larger loss rates.
During the last cycle, accelerating home price declines started a negative feedback loop. Homes stopped selling, or were sold at lower and lower prices. Pre-sales walked away, forfeiting their deposits. Commercial property developments beside unsold/incomplete home developments could not be rented, causing these projects to experience losses. Personal guarantees by developers were exhausted. This caused further declines in home prices, exacerbating the cycle. It is worth noting that in the U.S. even the best and most astute construction lenders, some of whom had never experienced any losses in their portfolios in a rising market, suffered.
Bottom Line: For most of the MICs we reviewed, historical losses were de minimis, even those with significant exposure to construction and development loans.
Lesson #4: Low Loan-to-Values May Not Provide the Protection One Might Expect
Another important potential issue for generalist investors is overestimating the protection (to both their investment and its yield) offered by “low” loan-to-values (LTV) ratios. For a MIC with a portfolio LTV of, say 60%, it does NOT mean that the portfolio will not realize any losses if property values decline less than 40%. In fact, LTV is often an imprecise metric.
First, many investors – incorrectly – believe “value” in the loan-to-value ratio always represents an independent number based on a variety of appraisals in a large liquid market (like houses in a neighbourhood). Sometimes it is supported by this data. However, in many instances, including development projects, the “V” is based upon, or heavily influenced by, a number of highly subjective assumptions including expected number of homes sold, prices of homes sold, and timing of sales.
Investors should be aware that most mortgages within a MIC are considered Level 3 assets according to the CICA Handbook, which means the following standard financial statement disclosure: “the Fund’s mortgage investments are measured at fair value using UN-observable inputs” [emphasis added]. Additional disclosure highlights that “by their nature, estimates of fair value are subjective …”. It is not clear to us that investors understand the degree of subjectivity incorporated into the estimate of “V”.
Second, the relationship between declines in property values and collateral values may not be linear, meaning a decline in home prices may not be the same as the decline in the value of the property. For example, if home prices decline by 20%, a land development project may see significant pre-sales walk away from their deposits, and may not end up not selling a single home before the developer runs out of cash, causing the collateral value of the project to decline dramatically more than 20%.
In the U.S., there were numerous instances where commercial or residential development projects saw 60% to 70% declines in market value. Raw or undeveloped land was even worse, where in extreme cases we saw recovery rates of only $0.10 to $0.15 on the dollar. While these examples were almost exclusive to local housing markets impacted by the worst excesses (i.e., parts of California, Arizona, Florida), the key point is that the ultimate recovery value for development projects in times of rapidly falling home prices can be much lower than the LTV at origination implies (because the “V” subsequently declined). Note virtually all LTV ratios are “at origination”.
Third, loan-to-value is an average and may obscure the actual credit risk of the portfolio. For example, a portfolio of 10 equal sized mortgages may have an average LTV of 60%, but may consist of 5 mortgages with LTVs of 40%, and 5 mortgages with LTVs of 80%. In this simplistic example, a 21% decline in property values could put half the portfolio underwater, and at risk of losses. This is, of course, a bigger issue for banks where the underlying portfolio is supported by a lot less capital.
Bottom Line: In most of the MICs we reviewed, especially the ones with outsized construction exposures, we believe their LTV would be significantly influenced by a large number of highly subjective management assumptions (i.e., consistent with CICA financial statement disclosure).
With Property Values Possibly Set to Decline, MIC Investors Should Exercise Caution
Should Canadian home prices decline materially, the consequences for all real estate lenders, including mortgage investment corporations, is likely to be negative as losses emerge; the question will be how much. This is especially true for construction and land development loans. If the correction is severe, we would expect the loss experience of MICs to vary widely, based on their portfolios’ specific exposures, LTVs and concentration.
The tremendous growth in MICs, supported by the flood of capital entering the market through large capital raises, also gives us pause, since it may have resulted in an expansion and degradation of credit quality (as certain lenders may have gone further and further down market to ensure they deployed all of their investor capital).
However, our biggest concern is that many investors that depend on yield might underestimate the credit risk of a MIC investment and overestimate the protection they receive from a “low” loan-to-value ratio. Rising property values make it very difficult to differentiate between higher quality and lower quality MIC portfolios. Falling home price will likely reveal to investors who the best underwriters are.
To better assess the risk, investors should look for MICs to provide a detailed accounting of their loan portfolios with information regarding principal outstanding, rate, loan type, term, and collateral ranking (i.e., first, or second). Anything less requires great caution. For those MICs which are not publicly traded, we would also carefully review liquidity provisions in the offering memorandum.
We do not know if these vehicles will experience reduced payouts, or loss of principal, individually or en masse. High risk loans often still end up repaying their full interest and principal. However, we do know that if Canadian property values and/or new home sales decline materially, the risk of losses in MICs, particularly those with exposure to construction and land development, will rise significantly.
MIC loan portfolios yield 8% or more in an ultra-low rate environment for a reason.
Higher credit risk.
 Unfortunately, the FDIC only began to break out quarterly losses for both construction and development loans in 2003
 In fact, a more conservative measure is “loan-to-cost”, which measures the loan value to the overall cost of the project. Despite seeing significant exposure to construction loans, we saw few instances where MICs disclosed “loan-to-cost” in any meaningful way.