Risk Arbitrage: Anatomy of the J&J-Centocor Deal

With Centocor's (CNTO) acquisition by Johnson & Johnson (JNJ) just days away from closing, you would reasonably expect Centocor's shares to be trading just a few pennies below the takeover price. According to the announcement of the deal July 21, J&J agreed to pay $61 per share in its stock for Centocor. So why is Centocor available for purchase at 59 11/16, a $1.31 discount to the stated $61 deal price? (Note that all prices are as of Friday's close.)

The answer is that the pricing period is working against the interests of Centocor's holders.

Many stock-for-stock mergers are structured to provide a fixed value for the target's stock -- in other words, a set per-share value for the deal. This contrasts with the Microsoft (MSFT) acquisition of Visio (VSIO), where the exchange ratio is fixed, and the per-share value is left to fluctuate.

The challenge in structuring every merger is finding a method to deliver that fixed value in a way that is equitable to both buyer and seller. Since a merger can take several months to close, that can be tricky. The mechanism most commonly used by the bankers is a pricing period -- over which the series of stock prices are averaged to determine the deal's exchange ratio.

The specifics of a pricing period are always in the proxy summarized in the first few pages and described at length in the merger agreement, which is always included in the proxy. My advice: Always read the proxy so that you understand the dynamics of the deal in its final stages, and why share prices move as they do.

In Centocor's case, the exchange ratio will be set using a 20-day pricing period ending today -- two days before this Wednesday's shareholder meeting to approve the merger. The daily closing prices of J&J over the 20 days are averaged, and then divided into $61 to establish the exchange ratio.

The only way you will end up with precisely $61 in value for your Centocor shares is if J&J's share price at the close of the merger is exactly the same as the 20-day average -- an unlikely occurrence. By my calculations, the average share price over the first 19 days of the pricing period is about 95 19/32. Should J&J close today at Friday's price of 94, the average would be about 95 7/16, and the resulting exchange ratio set at .6391. The current value of the deal is then .6391 multiplied by J&J's current share price of 94, or about 60 1/16, implying there is 38 cents of spread left compared to Centocor's 59 11/16 share price.

What's in It for Me?

As an arbitrager, I want to minimize my exposure to J&J's share price swings, while still ensuring that I pocket the spread. The way to do this is to calculate for each day of the pricing period the short ratio for each share of Centocor held long. Divide the target's share price by the stated deal value (of $61), and multiply this by the number of the target shares held long. Then, sell short 1/20th of this number of shares. Make this calculation, and short sale, for every day of the pricing period, and make your short sales as near to the close as possible. This enables you to closely track the average price, and to lock in the $61 value. Of course, if J&J were to rise over the pricing period, we will be short at a worse price than the average and will have left money on the table.

Granted, this may be impractical for a small holder, given the commission for each daily sale. It may be more practical to sell stock every few days instead, which probably won't greatly influence your outcome. Over time, the small slippages both in your favor and against will likely even out.

The Centocor deal is complicated by the existence of a collar, which dictates the minimum and maximum share ratios the target is entitled to. These come into play if the buyer's stock moves up or down by a large amount from the merger's announcement through its closing, and the average price falls outside of the collar's parameters. Collars are complex, but because they aren't relevant here -- J&J's share price will average comfortably within the deal's collar -- we'll revisit them in a future column.

Some Variations on the Theme

Pricing periods are designed to be fair to both the buyer and the seller: A longer pricing period removes the likelihood a stock price can be manipulated to influence the exchange ratio. (Yes, it has been done.) Twenty days is about the longest you'll see, but there is quite a bit of variety out there.

Sometimes, instead of using the closing price, the volume-weighted average price is used -- i.e., the share price weighted by the number of shares traded at each price level. While theoretically fairer, since it is much harder to influence than the closing price, it complicates hedging: You need to sell stock throughout the day in small pieces if you want to track the spread precisely.

It may also complicate the normal chain of deal events, where the pricing period is followed by shareholder's vote, which is followed by the deal closing. But sometimes the pricing period runs to the day of the deal's closing, which usually can't be determined in advance. Or, the closing may be many days after the shareholder meeting, particularly if regulatory approvals are pending. You can really be flying blind trying to hedge yourself in those circumstances -- the pricing period may be running without you realizing it.

My all-time favorite twist is when the company defines the average as 10 randomly chosen days out of 20. Some investment banker or lawyer dreamed this one up, figuring this will keep the arbs out of the deal if the hedging can be made difficult enough. But it's a shortsighted move, because all it really accomplishes is ensuring the target's price doesn't trade as high as it otherwise would -- due to the uncertainty -- and I don't think it really keeps any arbs away.

In the International Paper (IP)/Union Camp deal, the companies unwittingly created a casino in the last days of the deal. Arbs were frantically running statistical packages to simulate millions of possible draws to determine if Union Camp's stock price had a built-in edge based on the range and expected values of the simulations. It turns out it did, by at least a few pennies with minimal risk, and so tens of millions of shares traded the day before the selection draw to game this anomaly. I don't think this is what the CEOs had in mind when they put in this silly provision.

David Brail is the president and portfolio manager of Palestra Capital, a Manhattan-based hedge fund that focuses on risk arbitrage. Prior to founding Palestra, he was director of research at hedge fund Dickstein Partners for eight years.

At the time of publication, Palestra Capital was long Centocor and short Johnson & Johnson, although holdings can change at any time. Under no circumstances does the information in this column represent a recommendation to buy or sell stocks. Brail appreciates your feedback at [e-mail].

As originally published this story contained errors. Please see Corrections and Clarifications.