An exporter bidding for a foreign public tender commits before having sold anything. The bid bond, required with the offer, is a bank undertaking to pay a sum to the public buyer if the bidder withdraws its offer during its validity period or refuses to sign the contract it has won. It is the only guarantee in this family living entirely before the contract, and it has an economic feature that makes it poorly understood: it ties up bank lines and exposes one to a call on deals one will not win, that is, statistically, on most of them.
One must first see what the offer's validity period does to a bidder, because that is where the real losses sit. An offer is firm for ninety or one hundred and twenty days, sometimes longer, and during that period the price is fixed while the exporter's costs are not. A rise in raw materials, a currency move, the loss of production capacity, can make the offer ruinous before it is accepted. The bidder then faces a choice no training prepares it for: perform at a loss, or withdraw its offer and let its bid bond be called. Both must be quantified before being in the situation.
A practice compounding that exposure is added: extension of validity. A public buyer whose procedure runs late asks bidders to extend their offers, and correspondingly to extend their bonds. Refusing means leaving the competition, accepting means carrying an old price into new conditions. That mechanism closely resembles the extend-or-pay seen on advance payment guarantees, and it produces the same effect: the exporter stays bound by a decision it does not take. The difference is that here it can still refuse, at a cost that is the loss of the deal and not the loss of the sum.
Political risk then intervenes in a way peculiar to this phase. A procedure can be cancelled by a public decision, a change of government can cause awards to be reviewed, an administrative measure can make performance of the contract just won impossible. In those cases the bidder cannot sign, and the buyer can call the bond for refusal to sign although the cause lies with it or with its state. That is the classic abusive call, transposed to the bidding stage, and it is handled with the same weapons: abusive-call cover reimburses after payment and on proof, and the proof is assembled during the procedure, not after the call.
The cost of this phase must moreover be gauged, and it is not nil even when all goes well. Each bid bond consumes a bank line, tied up for the whole validity of the offer, and an exporter bidding on ten contracts simultaneously ties up ten times. That consumption limits its capacity to bond the contracts it wins, so that too broad a bidding policy can stop it financing its successes. Following up releases is therefore treasury work as much as administrative work: a bond not released after a lost award goes on weighing, and nobody on the buyer's side has any interest in releasing it.
Finally comes the handover from one guarantee to the next, the worst-managed moment in the sequence. The successful bidder must replace its bid bond with a performance guarantee, and often provide an advance payment guarantee. Those three undertakings should never coexist on the same contract, and yet they regularly coexist for weeks, because release of the first is not requested when the second is issued. The exporter then stands guarantor twice for the same risk, with two bank lines consumed and two possibilities of a call. The release request is written the day the performance guarantee is issued, not later.
The discipline comes to four points, all before the offer is filed. Quantify what withdrawal during validity would cost, and compare it with what performance at a fixed price would cost if costs rose. Set a bond amount and a validity period in proportion, knowing they are rarely negotiable but that one can decline to bid. Track releases as one tracks receivables, because these undertakings consume the capacity that will serve the contracts won. And request release of the bid bond the very day the performance guarantee is issued, failing which one guarantees twice.
A French manufacturer files an 18 million euro bid on a public tender in North Africa, with a bid bond of 2 percent, that is 360,000 euros, and an offer validity of one hundred and twenty days expiring on May 15. On April 20, the buyer requests a ninety-day extension, the procedure having been delayed. Meanwhile, steel prices have risen 19 percent, bringing the expected margin from 11 to 2 percent, and a competing contract has just taken the workshop capacity earmarked for this one. The manufacturer hesitates between extending, withdrawing, or extending with a revised price. How should it decide?
The decision is made in figures and not in principle, and the three branches do not cost the same. Withdrawing now exposes one to the bond being called: 360,000 euros lost, with nothing in return, and a public buyer that will remember at the next tender. Extending without saying anything means carrying a February price into April conditions, on a margin down from 11 to 2 percent, and above all without the workshop capacity required, which moves the risk from price to delay in performance and therefore onto the performance guarantee that will follow: it is the branch that looks most prudent and is the most dangerous. Extending with a revised price is possible only if the tender rules allow it, which is rare, and most often amounts to a disguised withdrawal. What must be quantified before deciding is therefore the cost of performing at a 2 percent margin without available capacity, subcontracting or slippage included, and compared with the 360,000 euros. If that cost exceeds the bond, withdrawal is the right decision and should be taken squarely, notified properly. Two steps accompany the decision in every case: check the bank line tied up for four months and what it is preventing elsewhere, since that consumption is already a cost borne; and document from now the procedural delay attributable to the buyer, because if a public decision later cancelled the tender, that material is what would establish the abusive character of a call.
- 01The bid bond lives entirely before the contract, and it exposes one to a call on deals one will not win.
- 02During the offer's validity the price is fixed and costs are not: withdrawing costs the bond, performing can cost more, and this is quantified.
- 03The extension request reproduces extend-or-pay: one stays bound by a decision one does not take, but one can still refuse.
- 04Each bid ties up a bank line, and too broad a bidding policy stops one bonding the contracts one wins.
- 05Release of the bid bond is requested the day the performance guarantee is issued, failing which one guarantees the same risk twice.