The SeniorCare Investor: Ventas Versus Kindred
Within a week of the Centers for Medicare and Medicaid Services (CMS) issuing its revised guidelines for changes in Medicare reimbursement for long-term acute care (LTAC) hospitals, Ventas (NYSE: VTR) sent its long-awaited (or dreaded) rent reset proposal notice to Kindred Healthcare (NYSE: KND) with an opening salvo that shocked a few industry participants, and it didn’t take long for Kindred to fire back. The rent reset option is a right that Ventas has had ever since Kindred came out of bankruptcy protection five years ago and it covers 186 skilled nursing facilities and 39 LTACs in four Master Leases.
The theory in 2001 was that Ventas would provide Kindred with some post-bankruptcy rent relief (about a $50 million decrease in its annual rent obligation) to enable KND to strengthen its financial condition in the first several years after emerging from Chapter 11. In return, in 2006 Ventas would have a one-time right to reset the rent, as well as the annual rent escalators, to "market rates," but in no event could the annual rent be decreased from current levels. Ventas also received 1.5 million shares of Kindred’s common stock, which would have been worth more than $60 million in the past year, but we believe VTR disposed of its Kindred stock in the past two years. The rent reset issue is now front and center for the two companies and their investors, so sit back and grab a cup of coffee (or glass of wine) and we will try to dissect the situation as we see it.
For about the past year, Ventas has publicly stated that they thought the "market rent" for the Kindred leases would be at least $35 million higher than current levels. Ventas shareholders understood this to mean a floor, and we assume the higher (future) cash flow soon became embedded in VTR’s share price. On the other side, additional annual rent of at least $35 million became embedded in KND’s valuation as well, and many REIT and provider analysts began using this number as a proxy for their valuation models, whether it was fair or not.
And then during the first quarter this year, CMS floated a trial balloon for proposed changes in LTAC Medicare reimbursement for short stay "outlier" patients, which alone could have reduced KND’s annual LTAC revenues by up to 11% and caused a severe shortfall in projected cash flow. Obviously, this presented Kindred with an opportunity to portray the rent reset as potentially a nonevent, because the profitability of the LTAC division would likely revert back to where it was five years ago (or worse) when the Ventas rents were reduced in the first place. It also put Ventas in a little bind, as it could not present its reset notice until CMS clarified the final reimbursement regulations. And Kindred was in the uncomfortable position of lobbying CMS for a much smaller cut in the rates, knowing that for each percentage point gain in revenues it might get back, there could be a not quite as large offsetting increase in rent charged as a result of VTR’s reset. Easy come, easy go, as they say, but Kindred had to take one step at a time, and minimizing the revenue reduction was goal number one, especially since it has other LTACs not owned by Ventas.
In early May CMS finally came out with its final rules, and while less painful than the trial balloon, KND reported that it would result in an approximately $46 million decrease in its LTAC Medicare revenues (remember, not all of its LTACs are owned by VTR). When other changes are added, including the change in the DRG re-weighting rules and the elimination of the market basket adjustment, KND stated that the total reduction to its LTAC Medicare revenues would be at least $125 million in one year (for all of its LTACs). What was not disclosed was where costs could be lowered, if any, and how the company would change its patient mix, which doesn’t happen overnight, to minimize the revenue and cash flow shortfall.
It didn’t take Ventas long to digest the final Medicare changes, and just one week later the REIT fired off its rent reset notice to Kindred. The proposed increase was a bit of a shock (and this is an understatement) to the market, as Ventas started the negotiations with a proposed annual rent increase of approximately $111 million, just a tad above the "at least $35 million" increase the market was expecting. This would represent a 54% increase in the rent paid to Ventas, and a 39% increase in KND’s overall rent obligation. For comparative purposes, the $50 million rent reduction five years ago represented about a 22% rent cut for Ventas and a 19% overall rent decrease for Kindred.
Kindred’s return volley was to state that its appraisers had determined that the current rents are already "well above fair market" and that even without an increase, "it may be financially and strategically beneficial for us not to renew one or more of these bundles of leased properties" that are part of one Master Lease that expires in 2008. Well, this certainly set the stage for some tough discussions.
Of course, Ventas investors were pleased, because anything above $35 million would be gravy in their minds. But investors are not that naive, and given that Ventas’ shares barely moved on the reset news, no one is expecting a $111 million rent increase. Similarly, Kindred’s shares dropped by just 6% but have gained most of that back in a generally negative market environment. The second part of the reset notice was the proposed change in the annual rent escalator to 3.0%, which represents a slight reduction from the 3.5% annual increases currently in place. We will get to that later.
Kindred and Ventas have 30 days, until June 8, to negotiate a new annual rent and escalator. If they don’t come to an agreement, they have until July 8 to each designate an appraiser, who in turn will have 10 days to agree on a third appraiser to actually do the valuation work. If they can’t agree, then the American Arbitration Association will appoint one, and that appraiser will have just 60 days to render an opinion. This process seems flawed, since Kindred has stated it has already used five independent appraisal firms in the past year or so to evaluate all the properties it currently leases from Ventas, and Ventas has also hired appraisal firms to do the same work. Since we can probably count on one hand the number of appraisal firms that are qualified (okay, maybe a few extra fingers too) and have the depth to do this volume of work in 60 days, who is left? Why wouldn’t any of the firms that have already done consulting work, for either company, be conflicted out of the process? The easy answer is that there wouldn’t be anyone left. And, who’s to say the one that gets picked will actually want to do the work and be in the hot seat of the industry. They may politely decline, or decide to double their rate, which would be the smart thing to do since we assume it won’t be competitively bid.
The appraisal process is flawed in another way. Nine out of 10 times (if not more), an appraiser is asked to value a property, not what rent a willing tenant would pay to a willing landlord. The subjectivity of a "market rental rate" is significant, and more so than the subjectivity of a full property valuation. Appraisers are not used to looking at what a market "coverage ratio" would be for leases, and when they look at "market" leases, do they really know what went on behind the scenes? For example, how often have we seen an aggressive lease (meaning low initial coverage) and thought, are they out of their minds? What we, and the appraiser, may not know is that a particular deal may be done with an existing tenant for relationship purposes, or perhaps there are other assets cross-collateralized to protect the landlord in case the low-coverage leases run into problems.
Alternatively, a high coverage lease could be done because the landlord was aggressive on some others for the same customer. It is the overall credit they are looking at in these instances, not the actual rentals. So are these "market" rentals and "market" lease coverage ratios? You don’t really know. That is one of the reasons we do not use sale/leaseback transactions in our acquisition market statistics when the lessee is also the seller, because the price might be higher with a lower coverage, or the price might be lower with a higher coverage, or there may be other guarantees involved driving the ultimate structure and price.
In the Ventas-Kindred case, Ventas derived its $111 million rent increase by assuming a skilled nursing facility EBITDAR-to-rent coverage of between 1.1x and 1.2x. Most REITs don’t disclose the rent coverage ratios on their deals anymore (and many don’t individually disclose deals unless they are material), but in looking back in our records the range recently was from 1.1x (yes, they do exist) to 1.8x, with the average more in the range of 1.3x to 1.4x. The main reason is risk— reimbursement risk, regulatory risk, liability risk, 35- year-old property risk—and most landlords want a bit more wiggle room when the intellectually-challenged residents of our nation’s capital decide it’s time to play with the rules (again). And once again, we don’t know if these are "market" leases.
We are, however, seeing initial coverage ratios between 1.0x and 1.2x for assisted and independent living facilities, where government reimbursement does not play a role and they can charge market rates to their private pay clientele, but these properties tend to have been built in the past 10 years on average, not in the 1960s and 1970s. And according to Ryan Beck & Co., as of the end of last year the average EBITDAR coverage ratio for the SNF portfolios of four other REITs was 1.6x, 1.6x, 1.5x and 1.3x, but these are mostly mature leases and not going-in coverages.
Regarding the annual rent escalators, the recent market seems to suggest 2.0% to 2.5% for nursing facilities. Ventas knows it has a good thing going with its current 3.5% annual increase, which is why it was willing to offer a slight decrease to 3.0%. But our guess is it is doubtful that any appraiser will come up with many recent comparables to justify that.
For the LTACs, Ventas has set EBITDAR-to-rent market coverage ratios at about 1.4x to 1.5x. Unfortunately, we don’t have enough information on LTAC lease rates to determine how realistic that range is, but we hear that 2.0x or higher is more in the market. The larger dispute, however, will center around what is being leased, the real estate, or the real estate and the business, where the business value often exceeds the "real estate" value with LTACs. Ventas believes this argument is largely irrelevant, while Kindred believes this is crucial to an appropriate valuation and lease rate on the real estate.
As part of our analysis, we decided to look at how Kindred is doing today compared with five years ago when it was given a break on its leases with Ventas to get back on its financial feet, using the first quarter results of both 2001 and 2006 (this is company-wide and not just the Ventas-owned assets). In the quarter five years ago, KND’s skilled nursing division had a net operating income (NOI, and before rent) of $70.5 million, which dropped 31% to $48.6 million in the first quarter this year. The LTAC division had a quarterly NOI of $54.8 million five years ago, which almost doubled to $104.0 million in the 2006 first quarter. Combined, the NOI increased 22% from $125.3 million in the first quarter of 2001 to $152.6 million in this year’s first quarter. The operating margin of the SNF division dropped significantly to 10.1% in the five-year time period, while the margin of the LTAC division increased from 20.1% to 24.2%.
So what does all this mean? If an NOI of $125.3 million was low enough to warrant a temporary rent reduction, it would seem that a $111 million annual rent increase on the Ventas-owned assets would put Kindred in a similar financial condition as when it emerged from bankruptcy protection five years ago, but without applying the negative impact of the recent Medicare reimbursement changes for LTACs. One would then have to question the reason for the rent break five years ago, but we don’t want to go there.
In addition, total rent in the first quarter five years ago (before the rent reduction) was about 60% of the total SNF/LTAC net operating income, which dropped to 46% within a year after the rent reduction. In the first quarter of 2006, total rent was 47% of the SNF/LTAC net operating income, which would go to 65% if the full $111 million went into effect. While we have just picked two points in time, and other quarters would probably look different, they represent the numbers for the last quarter before emerging from bankruptcy and the most recent quarter, which seem to have some degree of objective relevance.
There is another side of the argument, and this relates to the definition of fair market rental under the Master Leases. Apparently, the new lease rates are to be based on what a willing lessee would pay, and what a willing landlord would accept, in a transaction between unrelated parties. Using this definition, Ventas has made the argument (put in our simplistic terms) that regarding the skilled nursing facilities, Kindred’s operating margins are below where they should be, and another operator would be willing to pay a higher rent because they would have a higher cash flow. While we don’t know what Ventas would improve, it is probably true that a local or small regional operator would have significantly lower liability and workers’ compensation insurance costs. We have seen in some of Kindred’s divestitures that a breakeven nursing facility can become profitable the next day with a new owner just because of lower insurance costs, and in some states that can be a big swing. This does make us a little nervous on the "fair market" side of things, because you now qualify the type of buyer (small with lower insurance costs), as opposed to looking at the current operations, but appraisers do this all the time with costs. So from a purely objective perspective, we can understand why Ventas believes it should receive higher rents, but since Kindred is not a small operator with low insurance costs, we equally understand why its cost structure is not going to change soon and that it will be at a financial disadvantage with higher rents.
Getting back to the surprise $111 million annual rent increase that Ventas proposed, we have to assume it was a strategic negotiating ploy, and perhaps a brilliant one at that, because why start at $35 million when you can only go lower? Unfortunately, a $111 million increase would have dropped Kindred’s first quarter pre-tax income from $35 million to $7 million, which seems a tad harsh, especially for a company with over $1.0 billion in quarterly revenues. And think of the psychological damage done. The appraisers are thinking Ventas is at $111 million while Kindred has argued that the rents are already at market or above market today. Splitting the difference yields a $55 million increase, which is significantly above the "at least $35 million" used for the past year and would be a win for Ventas. We don’t think that will happen, but when the subjective (or tweaking) part of the appraisal process occurs, the stance taken by both sides will certainly be lingering in the back (if not the front) of the mind of the appraiser. In other words, the $35 million increase is already embedded in both companies’ stock valuations (or so we assume), and the appraiser probably knows it. Will Ventas management cry if they don’t get the $111 million increase? Hardly. Will Kindred management cry if it ends up between $35 million and $55 million? Most definitely.
At least three analysts have come out with "guesstimates" of where the new rents will be, and they are $20.5 million, a minimum of $30 million and a range of $25 million to $50 million. We have no inside information on the Ventas facilities, but our bet would be an increase between $20 and $35 million, with a rent escalator between 2.0% and 2.5%, not because that is where we think the numbers are (we really don’t know the details and were not asked for an opinion), but because we think the designated appraiser will feel the need to justify some increase. The reality, however, is that the stakes are just too high for both sides to go the appraiser route, and even though their starting positions appear to be too far apart to sit down at the table and hammer out an agreement with any degree of civility before June 8, that would be the least risky course for both sides.
One option that was discussed in the market last year, but one that we haven’t heard lately, is for Ventas to accept a lump-sum cash payment from Kindred, with no annual rent increases but with the escalators remaining above 3.0%. This would keep a large ("at least $35 million") annual rent increase out of Kindred’s future earnings stream in perpetuity, helping the stock valuation, and the cash payment would be a one-time charge, something that corporate America has grown quite fond of in recent years. For Ventas, it would have a large sum of cash to invest or pay down debt, and it would lock in an attractive growth rate on the assets operated by its largest tenant, something which would go a long way towards increasing VTR’s annual dividend rate every year. This would be face-saving for both sides as well, because it changes the nature of the negotiation, and no one loses. Unless they have already gone down that path, with just a week to go it is probably too late. Unless, of course, the appraiser comes up with a valuation and then does a net present value equivalent option for the two sides to masticate. Hmmm.
That being said, with all of the public opinions expressed in the past several months, there is little likelihood the designated appraiser would come up with a zero increase, and we give a $111 million increase a 0% chance of happening (although if we had to bet on one option or the other, we would take the zero increase, since the other is two to three times the highest estimates previously discussed in the market). But there is a breakeven point for Ventas between a smaller than hoped for annual increase with lower escalators on the one hand, and no rent increase but keeping the above-market 3.5% escalators on the other hand (which produces an attractive annual cash flow increase). The biggest issue will be the LTACs, because of their higher cash flow and the impact of the reimbursement changes, and because there is a larger potential discrepancy between what the real estate is "worth" and what rent a tenant would be willing to pay for the right to operate that real estate as an LTAC and obtain the business cash flow. Ventas and Kindred strongly disagree on this issue, but the interpretation can result in a difference in rent for the LTAC portfolio of millions of dollars. And that’s real money.