They beat the second-highest bid by PLN 21.4 million. Three years later they wrote down PLN 38.7 million

On Thursday at 7:20 p.m., two people were left in the meeting room on the fourth floor: the company’s CFO and me. She had two spreadsheets side by side on her screen. On the left was the model on which her company had based its acquisition of a competitor three years earlier. On the right were that competitor’s results for the last twelve months.

According to the model, EBITDA in the third year after the acquisition was to reach PLN 14.6 million. The right-hand spreadsheet showed PLN 9.1 million. The auditor had not accepted the assumptions on which the company based its impairment test of the goodwill from this acquisition. The new valuation for the test was prepared by the Capital One Advisers team. After the test, the company wrote down PLN 38.7 million of goodwill.

The CFO scrolled through the model row by row, looking for an error in the formulas. She found none.

Goodwill is the excess of the price over the fair value of the assets acquired less liabilities. Groups reporting under IFRS test it every year. Companies reporting under the Polish Accounting Act amortize goodwill and test it for impairment when there are indications of impairment, such as a marked fall in results.

I have changed the details so that neither the companies nor the people can be identified.

Three years earlier the competitor had been sold in an auction process. Seven non-binding offers came in during the first round, and three bidders went through to the second round. At the time the competitor had EBITDA of PLN 10.3 million. The auction was won by the company the CFO works for. It paid PLN 92.8 million, nine times EBITDA. That was PLN 21.4 million, or 30%, more than the second-highest bid of PLN 71.4 million. In the photo from the signing, everyone is smiling. So is the CFO.

Boards that have won an auction usually tell me they won because they understood the company best. I think that is a mistake. Winning an auction tells the buyer above all that it valued the company higher than everyone else did. Until it can show where that difference comes from, it should treat it as an error in its own valuation.

A jar of coins

Economists call this the winner’s curse. It was first described in 1971 by three engineers at the oil company Atlantic Richfield: Capen, Clapp and Campbell. In 1983 Max Bazerman and William Samuelson demonstrated it in a simple experiment. In microeconomics classes at Boston University, 419 MBA students bid in 48 auctions for jars of coins or paper clips. Each jar was worth $8, which the participants did not know.

On average the students valued the contents of a jar at $5.13, a cautious estimate. Yet the average winning bid was $10.01, and the winner lost $2.01 on average. In more than half of the auctions the winner paid more than the jar was worth. The group as a whole did not overvalue the jar. The auction went to whoever had made the biggest error on the high side.

Where real money is at stake, it looks much the same. Richard Thaler cites data from the 1969 auction of oil rights in Alaska. The winning bids there totaled $900 million, and the second-highest bids for the same tracts $370 million.

The curse grows with the number of bidders. Assume that each buyer values the company honestly and that the valuations are spread around its true value with a standard deviation of 10%. On that assumption, roughly two valuations in three differ from that value by no more than 10%. On average, the highest valuation is overstated by 5.6% with two bidders, by 10.3% with four, by 12.7% with six and by 15.4% with ten. The figures come from a simple model and serve only as an illustration. The direction is certain: the more buyers there are, the stronger the proof the winner needs that its edge comes from knowledge.

Expected value of the highest of n independent valuations, normally distributed around the company’s true value.

The same mechanism works wherever the most extreme valuation wins. In a tender decided on the lowest price, the contract goes to the contractor that underestimated its costs the most. A candidate courted by several firms is hired by the firm that rated him highest, even when it rated him too high.

Why an experienced board pays anyway

Arithmetic explains where the error comes from. An experienced board fails to correct it for reasons that lie outside the spreadsheet: in costs already incurred and in the rules by which the people working on the deal are assessed.

Due diligence costs the buyer anything from several hundred thousand to more than a million zloty. Once that money has been spent, withdrawing looks like a loss and raising the bid looks like saving the investment. The deal team has been working on the acquisition for months and is often assessed on deals brought to completion. The board has promised the supervisory board growth through acquisitions. A lost auction means going back to the supervisory board empty-handed.

The curse is already at work in the non-binding offer, the preliminary bid in which the buyer gives a price range before due diligence. A bidder that wants to go through to the next round has a reason to give a high range. In the second round that range becomes a reference point for the board, and it is hard to go below it.

In 2016 Eric de Bodt, Jean-Gabriel Cousin and Richard Roll analyzed 1,935 acquisitions of US listed companies made in 1994–2014. They linked overpaying to conflicts of interest on the buyers’ boards. They also found indirect evidence that part of the overpaying comes from underestimating the winner’s curse.

Then there are the advisers. In July I wrote that an adviser paid only on success cares above all about getting the deal done, and the price matters less to him. On the buyer’s side it works the same way: the larger part of the fees paid to its advisers is usually payable when the deal closes.

The seller contributes as well. Its adviser invites as many buyers as possible into the process, sets their deadlines and releases information in measured doses over successive rounds. I do the same when I advise a seller, and I regard it as honest work for the client. Such a process raises the price partly because it increases the number of valuations from which the seller picks the highest.

A price limit signed before the bids are opened

Two years ago I advised a family-owned industrial group that wanted to buy a company from a private equity fund. Four buyers went through to the second round. A week before the deadline for binding offers, my client’s board asked me for one number: the price above which the deal stops making sense. We set it at PLN 48.5 million. I wrote it on a sheet of paper, which all three board members signed. Going above that price required a new board resolution and a written justification.

On the day the bids were opened, the seller’s adviser let us know that our offer would not be enough. The CEO wanted to add more. I put the sheet of paper on the table. I admit that that evening I was not sure myself whether we had been too cautious.

We learned the winner’s price a few months later from its financial statements filed with the National Court Register (KRS). It paid PLN 55.2 million, 7.9 times the EBITDA we knew from the process. Our limit was roughly 6.9 times that figure. Fourteen months after losing the auction, the same group bought a comparable company for 5.8 times EBITDA.

A board that is submitting a binding offer in the coming weeks should set such a limit before its last conversation with the seller’s adviser. Once the bids are opened, it is too late. From then on, each further conversation only pushes up the price, while knowledge of the company does not grow.

At Capital One Advisers we set the limit by splitting the company’s value into three layers. The first is the value of the company on its own, that is, the cash flows it will generate without any contribution from the buyer. Every bidder estimates this part, and this is where the winner’s curse operates. Its valuation has to be marked down more heavily the more competitors there are and the less is known about the company.

The second layer consists of synergies that any buyer from the industry would achieve, such as joint purchasing or one head office instead of two. Competitors will count them too and hand part of them to the seller, so this layer has to be competed for. The limit here is the amount above which winning no longer gives the buyer any benefit.

The last layer covers synergies available only to this buyer, such as its own sales network abroad or a cost that only it avoids. This layer is worth keeping almost in full. If the buyer hands it to the seller, it will be paying in advance for its own work after the acquisition.

After winning, and before signing the agreement, the board should write down the sources of its advantage over the losing bidders: what it knows about the company that they did not know, and what only this buyer can do with it. The list must contain specifics, for example the buyer’s customers who will start ordering the acquired company’s products, or the buyer’s technology that will cut the company’s production costs. Without such a list, the difference between the winning bid and the second-highest bid is most likely a valuation error.

What this means for the seller

An owner selling a company looks at the same arithmetic from the other side of the table. For him, the highest bid is the best one only if the buyer holds it until the deal closes. A winner that outbid the others by 30% runs the greatest risk that due diligence will not confirm its assumptions. It will find out when the seller is already bound by exclusivity.

In September I wrote about exclusivity after a letter of intent and about the 5% threshold beyond which a proposed price cut releases the seller from exclusivity. When the winner’s lead is large, this clause protects the seller the most. When we advise a seller, we ask the winner, before exclusivity is granted, to show how it built its price: what EBITDA it assumed, how it calculated net debt and how much synergy is in the price. A bidder that cannot do this usually cuts its price only near the end of due diligence. At that stage the seller has the least bargaining power, because rejecting the cut means going back to the losing bidders with a company the winner has walked away from.

When the winner is right

The curse concerns the part of the value that all buyers estimate from the same information, that is, the company’s cash flows and margin. A buyer that has something the others do not have can pay more without making a mistake.

In August I wrote about four valuations of one company, ranging from PLN 60 million to PLN 144 million. The highest offer came from a strategic investor that, without this company, would have spent four years building the same thing or would not have built it at all. Its offer reflected a cost that the acquisition allowed it to avoid, and it was not a valuation error. At most, it may have handed the seller more of that value than it had to.

Experienced buyers can allow for the curse in their bids. Audra Boone and Harold Mulherin studied acquisitions of US listed companies and showed in 2008 that competition in the sale process had no statistically significant relationship with the stock returns of the acquiring companies. In my view, this is what a market looks like when buyers lower their bids more the more competitors they face. The company from the Thursday meeting did not make that adjustment.

The trouble is that the synergies that raise the price by 30% are set out in the model far less precisely than the cash flows of the company itself. The buyer pays for them in cash on the closing date, and the risk of not achieving them stays with the buyer. The model the CFO had on her screen assumed that from the second year costs would fall by PLN 2.9 million a year. The company achieved PLN 0.7 million. Synergies therefore explain PLN 2.2 million of the PLN 5.5 million gap between the model and actual results. The remaining PLN 3.3 million is the weaker performance of the company itself: without synergies, it generated EBITDA of PLN 8.4 million in the third year, less than in the year it was acquired.

A private equity fund that has no company in the industry does not count synergies at all. Its valuation is pushed up by assumptions of higher debt, faster growth or a higher multiple on exit.

The hardest decision concerns the second layer. The buyer has to give part of it to the seller, or it will lose. How much it can give depends on the number of bidders, on how well it knows the industry and on how much losing would cost it. Auditors carrying out impairment tests and banks financing acquisitions see the consequences of this decision years later, often more clearly than the advisers who were present when it was made. This limit cannot be calculated with a single formula, and it lies in a different place in every deal. The family-owned industrial group that signed it on a sheet of paper a week before the binding offer deadline lost that auction. The winner paid PLN 6.7 million more than the group’s limit.

Sources

• Max H. Bazerman, William F. Samuelson, “I Won the Auction But Don’t Want the Prize”, Journal of Conflict Resolution, 1983
• Richard H. Thaler, “Anomalies: The Winner’s Curse”, Journal of Economic Perspectives, 1988
• Audra L. Boone, J. Harold Mulherin, “Do auctions induce a winner’s curse? New evidence from the corporate takeover market”, Journal of Financial Economics, 2008
• Eric de Bodt, Jean-Gabriel Cousin, Richard Roll, “Empirical Evidence of Overbidding in M&A Contests”, Caltech, 2016