The commercial mortgage-backed securities rating process is not particularly hard to understand: Investment bankers put together the information on their offerings and send it to all four agencies. A preliminary analysis is done, and on the basis of that, two agencies are hired.

Although the processes are similar at the four New York-based rating agencies - Standard & Poor's, Moody's, Fitch and Duff & Phelps - in some cases each approaches the business differently. And the same can be said for the investment bankers.

Charlotte, N.C.-based First Union Corp. works with all four agencies and makes its selection based on preliminary credit levels, but typically looks to have one of Standard & Poor's or Moody's Investors Service and one of Duff & Phelps Credit Rating Co. or Fitch IBCA rate the deals.

"Investors want to see Moody's or S&P on a deal," says Barry Reiner, managing director of First Union's commercial real estate finance unit. "A lot of times their charter document requires one of the two." First Union has done three securitizations this year, with Moody's and S&P as the rating agencies on one, Fitch and Moody's on another, and Fitch and S&P on the third.

Kansas City, Mo.-based Midland Loan Services Inc., now part of PNC Bank, prefers to use all the agencies because as Harry Funk, chief credit officer, observes, "all four agencies take a different approach when underwriting a real estate transaction, which helps to have several different viewpoints when trying to structure around a deal issue. Also, given the fact that when it comes time to securitize our transaction we are not sure which agencies will give the final ratings."

Chase Manhattan's CMBS group tries to work with all four during the course of a year. "Over time, we feel like we understand what they are doing and understand their criteria," says Scott Davidson, head of Chase's CMBS banking & trading unit.

Last year, CMBS volume grew to a record $78 billion, a great deal of business for the rating agencies. This year's volume will not be as strong, perhaps in the $60 billion range, which is still a large sum, but much less than the year before. In addition, one year ago, there was a high percentage of deals that were securitized at $2 billion or $3 billion and sometimes three agencies were chosen. This year, deals are closer to $1 billion and just two agencies are selected. So, last year four agencies were looking for three spots and now, like in the game of musical chairs one chair has been taken away, there are four agencies looking for two spots.

Roping the 'lion's share' From a market-share standpoint, all these factors make for a highly competitive environment. According to Commercial Mortgage Alert, of the 48 CMBS deals in the first half of 1999, Moody's rated 43 deals (85%) worth $28.9 billion; Fitch rated 28 deals (55%) worth $19 billion; S&P rated 29 deals (50%) worth $17.1 billion; and Duff & Phelps rated 12 deals (28%) worth $9.6 billion. Over the history of CMBS, nearly every agency has boasted the lead market share at some point, but today, with an 85% market-share, it would seem Moody's is on about every deal.

"What you can conclude from the market-share listings is that Moody's clearly has lower subordination levels, which is why they get on the lion's share of the deals," explains Heidi Silverberg, group vice president at Duff & Phelps in New York.

"There is a controversy going on among the rating agencies," says one source at a commercial bank. "One way to evaluate who is perhaps more aggressive is the market share. Underwriters go to the rating agency that is going to give them the best level. Obviously, that is implied. So you need to look at who is getting the greatest share of the deals as to who is giving the lowest levels."

As current market leader, Moody's catches most of the heat about subordination levels, and it is sensitive to the criticism. "Some of our competitors tend to overblow the situation and say we are buying ratings, but the fact of the matter is, they are on a bunch of deals with us and our methodology is consistent from deal to deal," explains Tad Philipp, a managing director at Moody's in New York.

From an issuer's standpoint, what it really comes down to is the lowest credit support levels, says Gale Scott, a managing director at S&P and head of the real estate finance group. "There are no two ways about it; issuers are looking for low credit support from the rating agencies. If I have the lowest credit-support levels on the street, then I will have an 85% market share like Moody's."

When the agencies come up with rating levels on CMBS or structured transactions, says the commercial banker, there is not such a significant amount of historical data to base conclusions on so that an agency will always arrive at exactly the same conclusion. "It's hard for me to tell you that I think 27% subordination in AAA bonds is the arguably correct number," the banker says. "So if someone says 26% or 25%, no matter what factor they are using, they are making a lot of assumptions."

According to the same commercial banker, Moody's has taken the view that their previous model was too conservative and is now a little more aggressive, "and that comes with the business," the banker says. "But, it's hard if you are one of the other agencies. You say, 'oh gee, it is winning the business because it is being more aggressive and giving levels that are too good, and that is reprehensible.' The truth is, I don't know how you get comfortable that any level is the right level. It's like sending someone to the moon and having the exact physics' calculations about how you get the ship there."

In January, a First Union securitization carried both a Moody's and S&P rating, which Reiner says is very unusual. "I think an investor would tell you S&P and Moody's on every deal would be their perfect scenario. Obviously, we adapt to the investor's point of view, and what makes them happy is to have one of S&P or Moody's and then someone else."

Most securitizations usually do not use only Fitch and Duff & Phelps on a deal because most investors usually like to see either S&P or Moody's connected to a transaction.

Reiner adds, S&P has been on the sidelines for a number of significant deals this year, "so again, investors were happy to see S&P rating our deals since they are now viewed as the conservative force."

Things are changing, and maybe Moody's has not realized it yet, says Scott. "The industry is very competitive, very intense, yet you won't see S&P on a number of transactions," she says. "We were not there for a bunch of conduit-sized transactions that came to market during the first and into the second quarter. We were not there for a good 10 transactions because our credit levels were not low enough. That's fine with us because we made a conscious decision to do that."

Scott says the investor base has taken notice of S&P's actions and wondered what happened. "That's good, because there is a reason we are not on them," she says. "We made a conscious decision not be on them. I could have lowered my credit support and been on any of them, but I won't do that."

Philipp counters, "Investors have a very powerful vote with their money, so if we have an opinion that is valued in the marketplace and we are not on a deal, investors want to know why. I have had investors tell me if we are not on the deal that makes them a little bit nervous, and they will look for some other deal. We are getting picked by issuers because they think it helps them sell bonds better. We have the broadest following."

"There is a bit of a bake-off and issuers certainly do consider enhancement levels," Philipp adds. "But if the levels are close, they look at service and they look at credibility with the investors."

Every issuer analyzes all four rating agencies' subordination levels and the economic impact of those subordination levels before they choose the rating agencies, says Chase Manhattan's Davidson. "However, in my experience, in m ost cases, the four are very tightly bunched and the differences marginal. You may find one agency that has a more aggressive view of your transaction and one that has a less aggressive view. The other views are similar. You keep the one with the aggressive view, and don't keep the one with the less aggressive view.

"With the other two, all things being equal, you look at service; investor opinion; ease of working on a deal; and, if all that balances out, who have I worked with the least and maybe I should get them involved with my next deal," he adds.

Subordination level concerns Houston-based L.J. Melody & Co., a division of CB Richard Ellis, simply originates loans for conduits and does not do any aggregation. Nevertheless, subordination levels are still a concern to the company.

"We originate production on behalf of several conduits, and those conduits ultimately get rated by one of the four rating agencies," explains Brian Stoffers, executive vice president with L.J. Melody. "But we are directly affected to the extent that subordination levels reflect on the ultimate spread to the borrower. That is what we sell - the ultimate spread - so we like good subordination levels. We like tight spreads, and that results in better overall mortgage rates for the borrowers. So we have a vested interest in keeping the quality of our mortgages up because our lenders will get better loans."

Much to Stoffers' dismay, there is, as he says, "a lot of internal debate" going on with the rating agencies and the business of low subordination levels. "We have heard rumblings from the investment-grade buyers that they are discounting somewhat the ratings on CMBS, because, if you step back for a moment and look at the real estate cycle, it doesn't match. We see record low subordination levels come out at a point in the cycle where we are close to peak in terms of overall strengths. That being the case, there is some speculation that the agencies are overly aggressive, and, therefore, one of the excuses for the spread widening is that investors are discounting the rating agencies somewhat."

This does not please Stoffers at all because when spreads widen it affects his business.

"Because of the deal flow in 1999, there is less product to go around amongst the rating agencies, so perhaps they have to keep levels low if they want to stay in business and keep their staff employed," Stoffers speculates. "It's an element of, 'Let's stay active in this market. We have to bid on everything.'"

This apparently is S&P's point as well. "We don't believe credit support should have dropped as low as it did when you compare it to what was going on last year," says Scott. "When you look at the pools and put them next to each other, the only thing that has changed in the pools is that, from 1998 to 1999, they have gotten smaller. We are extremely consistent when it comes to credit support. If you just lay out all the deals we have rated one by one, you would see that we didn't understand low, and when compared to the credit support levels we gave in 1998, coming down five or six points on a AAA for the same issuer just did not warrant it."

In April, Scottsdale, Ariz.-based Finova Group Inc. did its first CMBS issue, a $760 million securitization. After submitting information to all agencies, it elected to use S&P and Fitch. David Rogers, managing director of Finova Realty Capital says there has been a lot of "scuttlebutt" being tossed about concerning low credit levels. He says, the issuer looks for lower levels, "But in the long-term interest of the CMBS market it's important that people believe the ratings. You don't want to get into a situation where the rating agencies are perceived to be giving credit levels that are less than they should be, as that will create a situation where investors won't totally believe the ratings."

He adds, "We are not in that situation today."

A key rating in any portion of a CMBS transaction, Rogers says, is that which goes to the AAA and then the other investment grades. "From the issuer's standpoint, a 1% difference in subordination levels doesn't kill the deal. The kind of volatility in spreads we see today just completely overshadows any impact on conduit profitability due to a minor adjustment that may occur in the rating agency support level. Whether you are at 29% or 30% AAA, levels have a lot less impact than how your deals turns out if you had been expecting pricing of 140-over for AAA, and then you watch what happens over the prior three weeks and now the deal is pricing at 160-over."

The ultimate clients are the investors, even though the issuer pays the salaries, Duff & Phelps' Silverberg stresses. "The only thing we really have to sell is our name and our reputation, and the goal is not to be on a deal and have the investor ask, 'why isn't someone else on that deal? Ultimately, you want the client to want you on the transaction, so, it's a fine line.

"Basically, what happens is we need to have different opinions," she continues. There are certain deals that we will like better than our competitors, so those are the deals we savor and we tend to get hired for. Each of the rating agencies have different things that are important to them."

The most important thing for issuers remains the levels, adds Janet Price, group managing director at Fitch. "We try to compete on other things because we can not compete on credit." She adds that sometimes agencies are selected on other criteria because they have a good investor base, service, accessibility of analysts, innovation and surveillance. "We hope people like those things we can offer, and they want to have us on a deal even if our levels are not the lowest," she says.

Taking different tacts All four rating agencies have a slightly different approach to underwriting. They each use guidelines that are slightly different in terms of arriving at underwriteable cash flow and value. "There are subordination models that are geared to different debt service coverage ratios," says Christopher Hoeffel, managing director at Bear Stearns in New York. "Although the agencies end up in the same place or close to each other, each has a particular sensitivity to different issues such as amortization cap rates, debt service coverage, loan-to-value, property types and individual components of underwriteable cash flow."

Most of the agencies would admit to this. As Fitch's Price notes, her agency has been very conservative on hotels, so it is not going to give the most favorable level on hotel deals.

"There is a certain amount of competition between the agencies," says Hoeffel, "and what I have noticed is that the agencies have picked firms they are comfortable with and for which they have generally rated deals where they like the collateral. If you are on one of those lists the agencies can be competitive in service, timing and fees."

Each rating agency takes a different approach to the business, Hoeffel continues. "There are rating agencies that want a lot of market-share and want to do a lot of deals. Other agencies say, 'I don't want to be on every deal; or I want to only be on every deal that this particular shop does; or I want to be on all deals that have this specific characteristic.'"

Some rating agencies will work on a flow basis with investment banks or other originators, Hoeffel says, "so they will give back on rating levels on an asset by asset basis. Obviously, they are doing that because they feel they have an inside track to getting the rating business and the deal securitized."

Midland has done one securitization so far this year for $750 million, and it was rated by Moody's and Duff & Phelps. "In order for Midland to receive better subordination levels it is important from the agencies' standpoint - in addition to Midland having good underwriting standards - that Midland have a well diversified pool of collateral from both a location and product standpoint," says Funk. "Therefore, because we focus on all of these items and our 'haircuts' have been around 3%, we feel that we historically have received good subordination levels from all the agencies."

The haircut? Funk explains: "If we as a conduit underwrote a particular deal that had $10 rents, 5% vacancy and a 4% management fee - and the agency concludes that market rents are actually $9, vacancy is 7% and their standard management fee for this product type is 5% - they would factor all of these differences into the deal and derive a haircut to our underwritten net cash flow."

Tricks of the trade There is no big secret about how agencies scrutinize a portfolio of loans. The investment bankers put together information that is then sent to all four rating agencies. Each agency does a preliminary analysis, and, as a result of that analysis, generally two agencies are hired.

Once an agency is hired, it picks a sample of loans to inspect and underwrite. That sample is typically 60% to 70% of the loan balance, but it is also a representative sample of the 10 largest loans by property type and geography. If there is more than one distributor of collateral, the sample would have to be the same for each party. For example, if there were three parties and they each had multifamily which accounted for 30% of the pool, then the rating agency would want to see 70% of each 30%.

After the inspection, the agency comes back and determines stabilized cash flow for each of the assets, and the results are extrapolated to non-sampled assets. The agency will look at what the average haircut was for multifamily for a given contributor and will use that average haircut to come up with stabilized cash for non-sampled assets.

Once the agency comes up with its calculations, the numbers are sent to the bankers and a conference call is set up to iron out the details. The rating process is very interactive, so the agency can change its opinions or simply agree to disagree with the issuer.

For example, the highlights of the rating methodology used by Fitch include:

* Originator Review. Originators are not rated, but rather, an assessment of the quality of underwriting for each transaction is incorporated into the subordination levels. Originators with experienced staff, training, formal approval processes, quality control, tight loan documents, quality borrowers and thorough income verification procedures will have lower losses over time.

* Sampling. A representative sample of the collateral by loan size, geographic location, property type, originator and other common features are reviewed in conjunction with the assets posing the largest risk to the deal.

* Site Inspections. Site inspections are performed to determine the quality of the property securing the loan and to verify the integrity of the data in the asset files.

* Reunderwriting. In reunderwriting a sample of a transactions' assets, the goal is to arrive at a steady-state net cash flow. Based on the underwritten sample, Fitch extrapolates the "haircut" (the percentage difference between the banker-underwritten net cash flow and Fitch's steady-state net cash flow).

* Modeling. Having derived a steady-state net cash flow for each asset in the transaction, a stressed debt service coverage is calculated by applying stress mortgage constants to calculate an adjusted debt service.

* Information distribution. The agency communicates to investors regarding its analysis with a presale report describing transaction strengths and weaknesses, important analytical considerations, collateral review and deal comparison charts.

At the end of the day... As Duff & Phelps' Silverberg notes, "There may be some biases, but none of the agencies charge for preliminary feedback. There is no incentive for bankers not to show everyone their deal because you never know what deal someone is going to like. We all have dealt with bankers so much we pretty well know where we are going to come out at the end of the day. We know the originator, we know the type of assets they lend on and we have a fairly good idea what kind of underwriting they have."

Ken Rivkin, head of mortgage-backed trading at Charlotte, N.C.-based Bank of America, notes that, while each agency has its own style, most of their methods are similar. "Their models are predicated on the impact of a collection of individual properties. They all look at a certain number of properties, and they all do some level of due diligence. There are some commonalties among them, but each one has its own separate model that weighs on different factors such as amortization, dispersion, property type and loan size concentration."

In a sense, any controversy involving subordination levels is not much different than that with the investment bank themselves, where the banks are often accused of reining in their analysts so as not to lose clients who might be offended by a poor analyst report. A poor analyst report might spur a large company to take its business elsewhere. Likewise, if the rating agencies promote credit-support levels that are too strict, a company might take its business elsewhere.

In reality, both the investment banks and the rating agencies would not be in business with credibility. Commerce is important, but in the financial services business, which includes the rating agencies, there would be no commerce without credibility.