Timbercreek Financial (TF) – a tenuous business model with hidden balance sheet risks (50% base case downside)

Portfolio Manager Summary (all amounts in CAD$)

  • We believe that the COVID-19 pandemic will accelerate the stresses already hidden in TF’s portfolio and expose TF’s underwriting and financing strategy
  • We believe that Timbercreek is grossly underprovisioned compared to its small-cap MIC peers – its current loan provision is 73% below the peer average – unusual to say the least. While its peers have taken actions to protect their balance sheets, TF has not, and we believe the dividend is at risk
  • In this report, we unveil two large exposures that collectively account for 14% of book value today – we believe they were significantly underprovisioned and that investors were inadequately informed about the issues at these properties – these assets were “sold” in a manner that, in our opinion, allowed TF to avoid taking appropriate provisions (our diligence, presented below, shows these “sales”)
  • These issues are critical in light of what we believe to be TF’s inability to sustain itself and its growth through internal cashflow generation. This calls into question the stability and security of the dividend given limited disclosure about the inherent risk embedded in their mortgage portfolio, as evidenced by the Sunrise portfolio and the Northumberland mall
  • We believe that this limited disclosure about underlying risks allows the company to retain the ability to obtain critical capital markets funding to support the dividend and support the stock in order to raise equity to (again) fund the dividend
  • Consider that old habits die hard – if TF is underprovisioned on two significant loans as our diligence suggests, we believe that is reflective of the tone at the top and likely throughout the organization. We infer that they must be materially underprovisioned elsewhere – COVID is likely to put TF’s book and provisions to the ultimate test and we believe the dividend is at riskin light of this, we assign a $4 price target to TF
  • Our summarized loan findings follow:
    • The Sunrise Properties’ loan appears to have been impaired since 2016 based on appraisals and bids received during the CCAA bankruptcy process that would imply material asset impairment
      • While TF no longer has a mortgage investment in Sunrise, it now owns the assets outright. After various attempts to sell the assets, TF had to abandon the sale process in late 2019 – these assets were likely impaired before COVID, thereby harming TF shareholders (the assets could only be worse after)
      • While Sunrise was initially a $28MM exposure to TF, management has since added almost $20MM to the investment as part of the credit bid and subsequent capital investment, avoiding a write down to their initial exposure
      • Since abandoning the sale process, management has been curiously quiet about the state of Sunset
    • The Northumberland Mall loan has been in the TF book since 2012 – the borrower went bankrupt in 2018. Despite 35% vacancy at Northumberland at foreclosure and plunging comparable mall valuations in the US, TF barely provisioned the exposure
      • We visited the mall in March, prior to COVID closures – it is despairingly vacant and stressed, as you can see in the photos below
      • TF appears to have used financial engineering to avoiding provisioning the mall – it “sold” the mall in 4Q19 by providing the “buyer” with over 100% financing – which brought the loan back into the “performing” category
      • With COVID further disrupting mall traffic, Northumberland is unlikely to be servicing its debt – we believe TF has limited tools left to avoid these losses
      • While initially a ~$35MM exposure, TF provided the mall “buyer” with a $55MM mortgage to potentially facilitate redevelopment and, we believe, to avoid a write-down

Executive Summary: Why COVID will expose TF’s hidden risks

Before the market sell-off, Timbercreek Financial (“TF”) appreciated over 37% in the last 5 years due to continue access to the capital markets and investor appetite for yield.  Unlike its Canadian counterparts who are down ~40% since COVID started, TF is down a mere 16% – we believe that this dislocation is unlikely to last (EQB – down ~37%, HCG – down ~41%, LB – down ~34%). Our diligence suggests that the impact of COVID will lay bare TF’s underwriting and put its dividend in jeopardy. We believe that TF is worth 53% below its current trading price, as we will detail what we believe are significant risks with filings and photographic evidence of their assets and loan portfolio.

In this article, we share with you our bespoke diligence which highlights two examples of significant loans made by Timbercreek (that reflect 14% of TF’s equity today) – we show how TF took what appear to be minimal provisions against clearly distressed loans and painted what we believe to be a rosy picture for investors. We believe that any provision taken on either asset would materially increase provisions while also constraining the company’s ability to fund its dividend through external financing. For example, we visited a mall 70 miles from Toronto that represents over 7% of TF’s book value – its vacancy is at least 25% and the property is in disrepair; we believe that COVID makes recovery here near impossible, yet TF provided over 100% financing to a buyer to move the loan into the “performing category” from the “impaired” category, likely making it wildly underprovisioned given the considerable risk. If we were able to find evidence for this for just two large loans, what does it say about the rest of the portfolio?

We believe that this, combined with a levered balance sheet and COVID-driven industry stresses, calls into question the stability of the dividend going forward, and we assign a $4 price target to TF’s stock.

Timbercreek Financial seems too good to be true

To begin, it’s important to understand why we started researching TF in the first place. TF’s February 2018 equity offering caught our eye – we noticed that TF was operating at 46.4% leverage, by their own definition, at 4Q17, implying that this equity offering was necessary to provide the funds required to continue growing the loan book (i.e., TF was unable to take on more debt). This led us to examine TF’s 4Q17 provision for mortgage losses, which we found to be just 9.7bps of the total $1.1B in mortgages, net of syndications, dramatically lower than its peers by a minimum of 82%. Put simply, the red flag investors should see here is that a company that is paying a healthy dividend should not be raising capital to fund that dividend over time.

A further examination of other small-cap Canadian mortgage investment corporations that lend into the residential, multi-residential, and commercial construction spaces shows that TF carries the lowest provision of the entire group, approximately 73% below the peer average of 87.9bps in 1Q20:

Worse, this seems to indicate that either TF is a superb underwriter of credit risk, or that it is not adequately provisioning for credit losses – the fact that the provision keeps rising would suggest the latter, and so does our deep diligence into two of TF’s large exposures. This is deeply concerning to us as most Canadian financial institutions are increasing loan-loss provisions. Despite indications of loan impairment and asset distress, we believe that the provisioning is inadequate and investors are unaware of the significant risks embedded in these exposures.

The Sunrise Portfolio – not as sunny as hoped

On TF’s 4Q16 conference call, then-CEO Andrew Jones mentioned a Saskatchewan investment that had sought protection from creditors through the Companies’ Creditors Arrangement Act (CCAA), Canada’s equivalent of the Chapter 11, restructuring process. 

A search of CCAA records revealed that this investment was most likely the Sunrise/Saskatoon Apartments partnership.

In late 2015, one of TF’s predecessor companies, along with other lenders, agreed to finance the acquisition and redevelopment by New Summit Partners of a set of apartment properties in Saskatchewan – of the 15 properties New Summit acquired, TF was the senior lender on 11 properties, which were known as the Sunrise properties. TF agreed to commit $115MM in funds for the transaction, and we believe they syndicated most of this amount, leaving it with ~$28MM of on-balance sheet exposure to the Sunrise properties, or approximately 4.3% of 4Q16 book value.

In June 2016, the Timbercreek entity associated with this loan became part of TF as we know it, through an amalgamation under the Business Corporations Act (Ontario).

Just six short months later, in December 2016, the Sunrise properties borrower applied for CCAA protection, citing “insufficient cash flow to complete development on the Properties in the face of demands for payment by their secured creditors and trades.” (unfortunately, the PWC CCAA page for the Sunrise properties was taken down sometime in the last few months, but the Government of Canada confirms PWC as the monitor and provides the same web address (now dead) where the documents we mention were sourced). You can try to obtain the records using the contact information found in the above link, which includes Michael Vermette at PWC, or by reaching out directly to the Vancouver Registry.

The CCAA process, which we show in the appendix to this article, resulted in appraisals and bids that showed that TF’s position as senior lender was in fact impaired, which we believe was not communicated to equity investors.

In the end, there was no market price for the Sunrise properties that would NOT impair TF and the other senior lenders’ positions in Sunrise, so TF ponied up additional capital to participate in the credit bid to acquire a 20% interest in 14 of the New Summit properties (including Sunrise), which increased their on-balance sheet exposure to $41MM (now $47MM).

In the 4Q16, 1Q17, and 2Q17 financial statements, TF repeatedly indicated that “there is no objective evidence of impairment”, which we believe contradicts evidence from the CCAA process – the following clippings are from the PWC monitor reports and/or affidavits taken during the CCAA process:

  • While appraisals were ongoing in 1Q17 and bids showed significant impairment in 2Q17:

While we believe that TF’s lack of provisioning (which it could have eventually reversed) in the face of numerous market indications of impairment was imprudent, and TF’s communications about this situation to its investors appear to contradict what was disclosed during the CCAA process.

Below we contrast comments made by TF’s founder and then CEO, Andrew Jones, to TF’s equity investors with statements found in the CCAA documentation: 

  • On TF’s 4Q16 earnings call, Jones said the properties had “…a sizeable amount of cash flow coming off…” them – but Tim Clark, New Summit’s GP, had the following to say in an affidavit dated 12/16/16
  • On TF’s 1Q17 earnings call, Jones said that the “renovation and lease up of those assets by the receiver or monitor is going very well” – but Jamie Dysart, Executive Director, Mortgage Investments, at fellow senior lender Kingsett stated the following in a May 2017 affidavit:
  • Lastly, on TF’s 2Q17 earnings call on 8/11/17, Jones said that TF was “controlling the asset and we have lots of equity” – but on 8/8/17, the lender group’s credit bid had been approved; by definition, a credit bid is equal to the value of the loan outstanding, implying that no equity existed. 

We believe that TF had to credit bid for these assets to avoid a significant impairment to their position as lender.

So what’s become of this portfolio? For several quarters after the credit bid, TF’s CEO Cameron Goodnough talked about selling the asset, as recently as the 2Q19 conference call:

But during the 3Q19 call, it looks like management abandoned the process:

This example is significant – at 4Q16, TF’s average loan size was just $8.2MM – the Sunrise exposure at this date was $27.6MM, or over 3x the company average. We believe that investors were largely unaware of the events and evidence we present here – evidence that could have significantly affected the company’s P&L to the detriment of its shareholders.

Our next example is similarly large but is a current loan exposure that materially deteriorated and could seriously affect TF’s financials.

The Northumberland Mall – can financial engineering hide bankruptcy forever?

In TF’s 2Q18 filings, a $36.9MM exposure move from Stage 2 (performing but increased credit risk) to Stage 3 (impaired) but with only a 1.2% provision:

1Q18

2Q18

On the 2Q18 earnings call in August, management reported to investors that the asset in question was current and in Toronto, per Director Ugo Bizzarri (but “near” Toronto, according to CFO Gigi Wong):

According to Teranet, on May 8, 2018 (in 2Q18), the mortgage for the Northumberland Shopping Centre, Inc., a mall in Cobourg, Ontario (population ~20k), was transferred to 2292912 Ontario, Inc, a subsidiary of TF, per TF’s 2017 Annual Information Form:

Not quite two weeks after the 2Q18 call, the Cobourg News Blog published an item saying Timbercreek had, in fact, taken over as owner of the mall:

And in fact, a Notice to Creditors filed on September 26, 2018 shows that the Northumberland Shopping Centre Inc, the borrower and owner of the Northumberland Mall, filed for bankruptcy, making TF the effective lender in possession of the Northumberland Mall:

Now why does this matter? Well, around the time that TF took possession of the asset, it was seriously distressed. Sandalwood, the property manager at the time, showed vacancy of approximately 130,973 square feet against leasable area of 370,769 square feet of leasable area, or occupancy of just 64.7%!

Even TripAdvisor visitors noted the level of disrepair the property was in:

We believe that given the high vacancy alone, TF should have provisioned more than just $463k on the property at the time. However, as the situation has become worse over time given the secular trend affecting malls and the ongoing despair at the Northumberland property, the provision needs to be materially higher, which will likely hurt earnings when this provision is appropriately taken.

In fact, a 2019 TripAdvisor post echoes the sentiment of the prior posts:

A new property manager, Trinity Group, was engaged last year to attempt to lease up the property, and has made progress – on January 17, 2020 the Cobourg News Blog posted an item outlining 95,966 square feet of vacancy against 375,000 square feet of leasable area (per Trinity Group), or an occupancy rate of 74.4%. It is also notable that as of February 3, 2020, Trinity Group removed the site plan shortly after the January 17th blog post.

If the stores looked like this before COVID, what might they look like after?

We visited the mall between the hours of 10am and 1:30pm on March 12, before the COVID closures.  This visit supports the conclusion that it still remains in distress and in need of capital – there is significant vacancy and lack of traffic:

Ocean Jewellers looks like it might just be opening late, but the mall directory shows it is no longer a tenant:

Even the existing stores are limited in staff, customers and size:

Metro:

Dollarama – a couple long empty aisles:

HR Fashion Plus

Hart

Sportcheck

Photos above indicate the emptiness and lack of staff this mall has in its stores. The outdated “entertainment” storefronts such as the cinema and bowling alley shows a low level of quality and care given to this piece of property.

The cinema has just three movies playing:

Mall employees said that 8 stores closed in the second half of 2019, and that every tenant in the mall is on month-to-month leases – which could heavily affect occupancy rate during COVID.

Given the level of vacancy and appearance of the property, we believe that investors may overestimate TF’s ability to “resolve” the asset:

2Q18, when the loan first moved into the Stage 3 category – this call occurred about 6 weeks before the borrower filed bankruptcy:

3Q18, after the bankruptcy filing, with guidance that the loan would be resolved by year end:

4Q18, guiding that the property will get “resolved” in the second quarter of 2019

On the 1Q19 call, no one mentioned the mall, but it was still on the books, and on the 2Q19 call, management claimed there were no questions AT ALL:

These clippings show that TF was telling investors it would resolve the asset, but nothing happened until over a year after foreclosing on the asset.

Well, so what’s happened to the Northumberland mall? According to Teranet filings, the mall was “sold” to Trinity Northumberland, Inc (presumably the same Trinity who was initially appointed as property manager) on 12/19/19 for $35MM, which happens to be pretty close to the carrying value of the outstanding loan on the balance sheet. At the same time, the purchaser took out a $55MM mortgage on the property, essentially financing 100% of the purchase and getting some more cash in the door for redevelopment purposes:

Now before we give you more details here, let’s talk about malls as an asset class. Much has been said about the demise of malls, most of it valid and relevant to this property. The Teranet data shows the initial mortgage being created in 2012, a time when lenders didn’t anticipate the kind of carnage that has ripped through B and C mall land (see NYSE:CBL and NYSE:WPG, for example) as a result of falling foot traffic and ecommerce. Today, lenders are underwriting B and C mall assets at upwards of 15% cap rates – but for a mall with a 25% vacancy rate and in sore need of capital investment, there is no telling what kind of valuation it could fetch.

In fact, similarly situated properties in the US have seen their loans massively impaired at maturity:

  • Fashion Outlets of Las Vegas, a now REO asset in Primm, NV, defaulted on its mortgage in 2017 with a year end occupancy of approximately 75%. It appraised for $125MM in June of 2012, when a $73MM mortgage was put in place, but the value of the property has since fallen to $28.8MM as of June 2019. The current appraisal value implies a 60% loan loss to the mortgage
  • Salem Center, a Salem, OR mall also defaulted on its mortgage in 2017 with a year end occupancy of approximately 79%. It appraised for $18.5MM in June of 2019, compared to $44MM when its $33.3MM mortgage was originated in 2012. The current appraisal value implies a 44% loan loss to the mortgage

Knowing this, how could TF justify financing 100% of the purchase price of a mall asset, and then throw another 57% of the purchase in as additional cash?

When asked about amount of the new mortgage on the 4Q19 call, Managing Director Scott Rowland simply didn’t answer the question:

So, dear shareholder, TF has “sold” an asset (and moved the loan backing it back from Stage 3 impaired to Stage 1 performing) by providing the lucky buyer with 100% financing (and then some)! We believe that TF likely rolled their loan outstanding into a new mortgage and threw in another $20MM in for the borrower to redevelop the asset, allowing the buyer to take ZERO equity risk in the transaction. Who wouldn’t want to buy an asset with no skin in the game?

Balance sheet risks are compounded by TF’s financing strategy

So why do Sunrise and Northumberland even matter?

  • They are/were disproportionately large exposures–the Sunrise properties’ loan balance of $28MM (4.3% of book value at 4Q16 and 6.6% of BV at 1Q20) was almost 3x TF’s average mortgage investment size, and Northumberland’s new loan balance of $55MM (7.6% of book value at 1Q20) is 5.8x the size of the average TF mortgage investment size
  • It was clear from appraisals and bids that TF’s position as senior lender to Sunrise was impaired, but investors were unaware of this
  • Given the state of the Northumberland asset and comparable transactions, TF’s position as a lender is likely materially impaired, but management provided a buyer with over 100% financing to take the asset out of an impaired category
  • What does this say about other loans that might face the same issues or fate at Timbercreek?

We believe that TF has been, and continues to be, dangerously under-provisioned and that as a result investors have been/are unaware of the underlying risks on the balance sheet. This reminds us of Home Capital (TSX: HCG) in 2017, which had a perceived high quality loan book but was subsequently exposed as hosting dubious mortgages with increased risk that were never disclosed to the market, reflecting >10% of their book value; the stock then fell over 90% reflecting these undisclosed issues. We are concerned that similar patterns could be taking place at Timbercreek Financial, and that TF is behind its peers in preserving capital flexibility:

These issues we’ve highlighted here are critical, as the TF bull case is largely based on Timbercreek’s perceived expertise and continued dividend payments. In addition to issues with individual loans, TF’s financing strategy poses significant risk to the dividend. We believe that TF doesn’t generate enough cash to service its dividend and grow the loan book, and that it is essentially borrowing money to pay the dividend. This represents an outsized and underestimated risks to shareholders, especially in light of TF’s already high payout ratio and tendency to roll a significant amount of maturities every year, coupled with the fact that, according to Trepp, CMBS mortgage delinquency rates are approaching all-time highs.

With the exception of 2019, in the fourth quarter of the prior three fiscal years, TF’s borrowers have missed over 40% of the scheduled maturities for that quarter, implying that TF had been regularly extending loan maturities at year end:

Practically, this dynamic means that TF is not receiving enough capital back at the end of each year to grow the loan book and service the dividend without external financing. In fact, over the last 21 quarters (1Q15 to 1Q20), TF’s operating and investing outflows have exceeded its operating and investing inflows by $71MM: 

The obvious implication here is that since operating and investing inflows are less than outflows, TF is actually unable to service the dividend without external financing – in fact, in order to plug the operating and investing deficit and address the buyback, interest payments, and service the dividend, TF has had to find over $403MM in external financing over the last 21 quarters:

Sure enough, over the same period, TF has drawn $409MM from the capital markets:

In addition to the dividend risk embedded in this sort of unsustainable model, we believe that an appropriate level of provisioning would result in a much lower stock price. In a normal scenario, we’d expect TF to appropriately provision for the Northumberland mall based on a combination of its peers’ provisioning and a loan loss severity for Northumberland consistent with the previously mentioned Fashion Outlets and Salem Center.

Things are hardly normal now though – Northumberland is a distressed asset whose traffic is now likely near zero, in a city whose population overindexes to older people, the demographic most seriously affected by COVID. We believe that the likelihood that the property ever really recovers is low, and that TF’s debt is materially impaired, if not entirely so. The ability of the mall to service the new $55MM note has likely evaporated, as news of malls tenants in other properties choosing not to pay rent abounds. We believe that a more reasonable loss severity for the Northumberland Mall is closer to 70% vs. the average 52% from Fashion Outlets and Salem Center, given that the asset and loan are now likely teetering on non-performance and the likelihood of turnaround is low. Assuming a 70% loss to Northumberland, we get to an average price target near $4, down 53% from the 7/28 close:

We believe that given the reliance on external funding, TF has to choose between growing the loan book or maintaining the dividend – unfortunately, given what we believe to be less than ideal capital recycling, TF cannot do the latter without the former. The Northumberland exposure, in addition to broad portfolio distress due to COVID, makes a dividend cut over the next 4 quarters a frighteningly likely outcome.

Wrapping it all up

In addition to what appears to be a risky financing strategy, we believe that the shareholder base is unaware of the risks we’ve outlined here. In fact, we are willing to bet that shareholders aren’t aware that TF provided a $100MM mortgage to Cresford Developments, a real estate developer who is subject of a lawsuit alleging a “cash crisis”and the misrepresentation of financial condition (Anthanasoulis v. Cresford; CV-20-00634836-0000; Ontario Superior Court of Justice):

Sure enough, in late March 2020, three of Cresford’s projects entered receivership – while the property for which TF provided capital was NOT impaired and included in receivership, the receiver’s findings included “evidence of a number of questionable accounting and management practices by the Cresford Group that had the effect of hiding substantial cost overruns on all the Cresford Projects [the projects in receivership]”.

These findings related to a TF loan recipient call into question the level of diligence and prudence with which TF underwrites its investments:

With one significant finding that could eventually affect TF’s own position:

In the end, we expect shareholders may be hurt by these dynamics and we hope that this article helps to illuminate the apparent risks hidden in this business.

Appendix – Excerpts from the Sunrise CCAA process

At 12/31/16, we see that TF is the largest creditor to the properties with $104MM of exposure, of which we estimate $28MM was on TF’s balance sheet, while the remainder was syndicated:

Thus begins a process by which TF received numerous market-based indications that its loan was impaired, but did not take a provision to reflect this impairment:

  • On April 4, 2017, TF and Kingsett (another senior lender) received property appraisals from Altus, which revealed that at current market values, the First Secured Lenders would be in a shortfall position:
  • In fact, the appraisals show the current market value of the Sunrise properties at $101.6MM against first lien debt of $103.3MM, an approximate 2% impairment to TF’s position, but no provision was taken to anticipate this loss in value
  • As the CCAA process continued, the lenders sought to sell the properties in groups – in May of 2017, a discussion was held concerning offers received for 5 properties known as the Group 1 Lands, showing that the properties with TF mortgages were significantly impaired

In June 2017, senior lender Kingsett sought the approval of a credit bid for the Group 1 properties and indicated it would also seek approval of a credit bid for the remaining nine Group 2 Sunrise properties. On June 16, 2017, the Court approved the sale of the Group 1 and Group 3 properties to various entities controlled by Timbercreek and Kingsett and ordered that the Group 2 properties (the remaining 9 Sunrise properties) be listed with Kingsett’s credit bid serving as the floor price for the process.

  • Unfortunately for the lenders:
  • But on 7/26, Bondstreet, who had only conducted limited due diligence, approached with a $131.5MM bid for these 9 properties, which it subsequently reduced to $115MM, below the value of Kingsett credit bid, thus implying an impairment to the creditor’s outstanding loans