European Construction Insurance: Bauleistung vs Tous Risques vs CAR Compared

Three policies that brokers routinely treat as interchangeable are not interchangeable, and the gap between them is where uninsured loss lives.

The German Bauleistungsversicherung, the French Tous Risques Chantier, and the English-language Contractors All Risks policy are sold across the same European market under the same shorthand of "all risks." They are written in different legal traditions, they default to different exclusions, and they treat the same site loss in materially different ways. Operators who run projects across borders inherit whichever form the local broker prefers, and the broker's preference is rarely a coverage decision. It is an administrative one. The contractor signs, the project starts, the loss happens, and only at the adjustment stage does the difference become visible. By then the form cannot be changed.

This article compares the three policy families from the perspective of an operator who buys insurance to transfer risk, not to satisfy a procurement checkbox. It draws on observations from sites in Germany, France, the United Kingdom and the Benelux region, on the standard wordings of the major European underwriters, and on the practical reality that what is excluded in one tradition is silently included in another. The argument is not academic. The amounts at stake on a single mid-size project routinely exceed the entire annual premium budget of the contractor's security and risk function. The book BOSWAU + KNAUER. From Building to Security Technology describes the construction site as an honest teacher. Insurance is the part of the construction economy where that teaching is most expensive when ignored.

The German Bauleistungsversicherung as a benefit-for-the-builder construction

The German Bauleistung is structured around a logic that the English market does not share. It is conceived as a benefit running primarily to the principal, the Bauherr, with the contractor and subcontractors named or implicitly included depending on the wording. The policy attaches to the construction work itself, not to a defined contractor's interest, and it pays for damage to the works during the construction phase. This sounds similar to a Contractors All Risks form, and at the level of the marketing brochure it is. At the level of the small print it is not.

The Bauleistung wording is narrower than its name suggests. It covers unforeseen damage to the works, but it traditionally excludes a long list of named perils that the contractor often assumes are inside. Damage from normal weather, including rain that any competent site planner should have anticipated, sits outside the cover. Damage to materials before they are incorporated into the works can fall outside depending on the wording. Theft is covered only where the material has been physically built in, which leaves the lay-down yard, the most exposed area of any site, dependent on a separate Bauleistung extension or on a stand-alone theft cover that many contractors do not buy because they assume their main policy handles it. The German market has refined these exclusions over decades of case law, and the courts have been consistent. What is not explicitly included is not covered, regardless of what the broker said at placement.

The strength of the Bauleistung is its alignment with the German construction code, the VOB, and with the contractual obligations the contractor owes the principal. When the policy responds, it responds in a framework that German lawyers, adjusters and engineers understand. The weakness is the same alignment. A contractor used to working under FIDIC, under English common law, or under French civil code provisions will find that the Bauleistung does not cover risks that those other frameworks treat as routine. The policy is precise where the German tradition is precise, and silent where the German tradition leaves the question to the underlying contract. A site that loses cable to theft in the third week, before the cable has been pulled into conduit, often discovers that the Bauleistung treats the loss as a stored-material question rather than a works question. The contractor pays.

The French Tous Risques Chantier and the civil code as silent partner

The French Tous Risques Chantier, the TRC, operates inside a different legal architecture. France runs construction risk through a system that combines the TRC at the project level with the décennale, the ten-year liability cover that follows the building after handover, and with the Dommages-Ouvrage, the principal's parallel cover that pre-funds repairs while liability is established. The TRC sits inside this architecture as the construction-phase piece, and its wording reflects that. It is broader than the German Bauleistung in some respects and narrower in others, and a contractor reading it for the first time should not assume that the French market's reputation for comprehensive cover translates directly into the policy text.

The TRC typically covers the works, materials on site, temporary works, and damage caused by the contractor to existing structures, the latter often through a specific extension. It tends to handle theft more generously than the Bauleistung, accepting that materials in transit to the site and stored within a defined perimeter are part of the project risk. It also tends to incorporate consequential exclusions that the German market handles through separate policies. Delay in start-up, advanced loss of profits, contractual penalties for late delivery, these sit outside the standard TRC and require named extensions if the contractor wants them inside.

The defining feature of the TRC, however, is what surrounds it rather than what it contains. The French construction code imposes obligations on every actor in the chain, and the TRC is calibrated against those obligations. The policy assumes that the principal carries Dommages-Ouvrage, that the contractor carries décennale, and that the subcontractors carry their own responsabilité civile. Where any of these assumptions fails, the TRC behaves in ways that surprise contractors from jurisdictions where insurance is the primary risk-transfer mechanism. The policy does not fill gaps in the legal architecture. It complements them. A foreign contractor entering the French market with a TRC and no understanding of the décennale obligation has bought a policy that covers the site phase and left a ten-year liability open. The premium looked competitive. The exposure was not.

The English Contractors All Risks and the appearance of breadth

The Contractors All Risks form, the CAR, originated in the London market and travels with the English-language construction industry across the Commonwealth, the Gulf, and most of the projects that involve international financing. It is written on the principle of named exclusions rather than named perils, which means that anything not explicitly carved out is, in principle, covered. This is the broadest of the three forms in its drafting logic, and it is the form that most international contractors default to when they have a choice.

The breadth of the CAR is real but qualified. The standard exclusions are extensive and have grown over the last two decades. Defective design, defective workmanship, defective materials, wear and tear, gradual deterioration, mechanical or electrical breakdown of plant, cyber risk, war and political risk, communicable disease, and an expanding list of consequential losses all sit outside the standard form. The list of exclusions in a modern CAR runs to several pages, and each line has been written into the policy because a loss has been paid that the underwriter did not intend to pay. Reading the exclusions is the only honest way to read a CAR. The marketing material describes what the policy covers in general terms. The exclusions describe what it covers in fact.

The CAR also carries deductibles and limits that operate differently from the European forms. Each-and-every-loss deductibles can erode a contractor's recovery on a series of small theft incidents to the point where the policy is, economically, a catastrophe cover with a long list of small uninsured losses underneath. The annual aggregate limits on certain extensions, theft and vandalism in particular, can be exhausted on a single bad month, leaving the contractor exposed for the remainder of the project. The CAR's flexibility is real for sophisticated buyers who negotiate the wording. It is illusory for contractors who accept the standard form and discover at the first claim that the policy they bought is not the policy they thought they bought. NIST CSF 2.0 has begun to influence the cyber exclusions in CAR policies, with underwriters tightening the language around electronic exposure as construction sites digitalize. The contractor who installs networked site equipment without checking the cyber exclusion in the CAR is repeating the pattern that produced the standard exclusions in the first place.

Coverage gaps that none of the three forms close

Across all three policy families there are exposures that the standard form does not address, and that are addressed only by extensions, separate policies, or contractual risk allocation between principal and contractor. These gaps are not accidental. They are the places where the insurance industry has decided that the risk is either uninsurable, too volatile to price, or better borne by the party that controls it.

Theft of materials before incorporation into the works is the most common gap and the one that produces the most frequent uninsured loss on construction sites. Bauleistung, TRC and CAR all treat stored materials differently from incorporated materials, and all of them impose security conditions, value limits and reporting obligations on the stored-material extension. A contractor who fails to maintain the security measures stipulated in the policy, lighting, fencing, guarding, surveillance, can find the theft cover voided after the loss. The IEC 62443 framework for industrial security, although developed for operational technology, is increasingly cited by underwriters as the benchmark for what reasonable security looks like on a construction site that incorporates electronic systems. The German GDV has published guidance that overlaps with this trend. The contractor who can demonstrate compliance with a recognized framework has a stronger position at the claims stage than the contractor who relies on traditional measures alone.

Delay-in-start-up and advanced loss of profits sit outside all three standard forms and require separate cover. The premium for these covers is high because the exposure is, in many projects, larger than the physical damage exposure itself. A six-month delay on a logistics facility can cost the principal more than the entire construction value of the building. Contractors who price these covers into their tenders and principals who require them in the contract are operating in a different risk class from those who do not.

Cyber exposure on construction sites is the newest gap and the one growing fastest. Site offices run networks. Site equipment runs firmware. Building management systems are commissioned before handover. A ransomware event during commissioning can halt the project for weeks. The standard CAR excludes it. The standard Bauleistung excludes it. The standard TRC excludes it. CISA advisories on construction-sector cyber risk and the ISO 27001 framework for information security management are both becoming reference points in the negotiation of these extensions, but the market is not yet mature. Contractors operating in this gap are bearing the risk whether they know it or not.

How brokers compare the three forms and why their comparisons mislead

A broker comparing Bauleistung, TRC and CAR for a multi-jurisdiction project produces a matrix. The matrix lists perils down the left, policies across the top, and ticks or crosses in the cells. The matrix is a useful starting point. It is a misleading conclusion. The reason is that the three policies operate in different legal and contractual environments, and a tick in one column does not mean the same thing as a tick in another. The Bauleistung tick for theft is qualified by the German market's narrow definition of when material is part of the works. The TRC tick is qualified by the French requirement that the principal carry Dommages-Ouvrage. The CAR tick is qualified by the deductible and limit structure that determines what the contractor actually recovers.

A serious comparison runs at the level of the claims process, not the policy schedule. It asks how each policy behaves under three scenarios that any contractor will encounter. The theft scenario, in which lay-down-yard cable disappears in the third week. The water scenario, in which an unusual but not unforeseeable rainfall floods the basement excavation. The third-party-damage scenario, in which the contractor damages a neighbouring building during piling work. Each policy handles these differently, and the differences appear not in the inclusion list but in the exclusion list, the deductible structure, the warranty conditions, and the interaction with the surrounding legal framework.

The broker who can walk an operator through these three scenarios in detail, with reference to the specific wording proposed and to the case law that has interpreted it, is doing the work that justifies the commission. The broker who produces a comparison matrix and a one-page summary is not. ASIS International has published guidance on how to evaluate construction insurance from a security perspective, and the better European brokers are moving in this direction. The operator who insists on this depth of analysis at placement is buying a different product from the operator who accepts the standard offer. NIST 800-53 control families are increasingly referenced in the placement of CAR policies for projects that include critical infrastructure, and the operator who can map their site security posture against a recognized control framework is the operator who pays the lower premium and recovers more at the claim.

What the three forms imply for cross-border construction operators

A contractor running projects in Germany, France and the United Kingdom in the same year is buying three different products under three labels that sound similar. The administrative temptation is to treat them as one product with local variations. The risk-management reality is that they are three products with overlapping but distinct coverage profiles, and the operator who treats them as one will at some point pay for the simplification.

The practical implication is that the risk transfer strategy of a cross-border operator cannot be built at the policy level alone. It must be built at the level of the underlying risk, with the policy chosen to address the residual after the operator has done what the operator can do. Security, surveillance, access control, materials tracking, all of these reduce the exposure that the policy must absorb, and all of them improve the position at placement and at claim. The ISO 27001 and IEC 62443 frameworks, the CISA guidance, the BSI publications and the GDV statistics all point in the same direction. The insurance market rewards operators who can demonstrate that they have done the work. It penalizes operators who buy the policy and treat security as the insurer's problem.

This is the point at which the construction-insurance question stops being an insurance question and becomes an operational question. The form of the policy matters. The exclusions matter. The deductibles matter. But the variable that determines the economic outcome over a portfolio of projects is the operator's ability to reduce the loss frequency that the policy is asked to cover. An operator running a site with documented surveillance, controlled access, recorded materials movements and integrated site security technology is, from the underwriter's perspective, a different risk from the operator who runs the same project with a fence and a night guard. The premium reflects this. The recovery at claim reflects this. The cumulative effect over a portfolio of projects, in any of the three policy traditions, is the difference between a construction business that absorbs its loss profile and one that pays it forward into next year's tender margin.

What holds

The three European construction insurance forms, Bauleistung, Tous Risques Chantier, and Contractors All Risks, are not interchangeable products with different names. They are different products with overlapping names, written in different legal traditions, calibrated against different surrounding obligations, and behaving differently at the claims stage. The operator who treats them as interchangeable is the operator who discovers the difference at the wrong moment.

The form of the policy is the start of the analysis, not the end. The exclusions, the deductibles, the warranty conditions, and the interaction with the surrounding legal framework determine what the policy actually covers. The operational measures the contractor takes on site determine what the policy is asked to cover. Both halves of the equation are within the operator's control. Neither is within the broker's control, which is why the broker's comparison matrix is the start of the conversation and not the answer to it.

For operators reviewing their construction insurance position across multiple jurisdictions, the first step is a structured conversation with someone who has read the specific wordings and walked the relevant sites. Path I in the engagement model described in the book BOSWAU + KNAUER. From Building to Security Technology, a sixty-minute confidential conversation, is the appropriate format for that step. It produces an assessment that places the policy question inside the broader risk-transfer question and identifies whether the next step is an audit of the site security posture or a direct conversation with the broker about the specific wordings in force.

Frequently asked questions

How do the three forms differ?

The German Bauleistung is a narrow named-perils policy aligned with the VOB and the German construction code, with traditional exclusions for weather, stored materials and theft outside specific extensions. The French Tous Risques Chantier is a broader policy that operates inside the French architecture of décennale and Dommages-Ouvrage, complementing those covers rather than replacing them. The English Contractors All Risks form is the broadest in drafting logic, working on named exclusions, but carries extensive deductibles, limits and exclusions that qualify the headline breadth. Each form behaves differently at the claims stage.

Which is broadest?

The Contractors All Risks form is broadest in its drafting principle, since anything not explicitly excluded is, in principle, covered. In practice, the exclusion list in a modern CAR runs to several pages and the deductible structure can erode recovery substantially on small and medium losses. The Tous Risques Chantier can be broader than the CAR for stored materials and certain third-party scenarios, depending on extensions. The Bauleistung is narrowest in its standard form. Headline breadth is not the same as effective recovery, which depends on exclusions, deductibles and warranty conditions.

Where does coverage gap?

Common gaps across all three forms include theft of materials before incorporation into the works, delay in start-up and advanced loss of profits, cyber exposure on networked site equipment, damage from foreseeable weather events, and contractual penalties for late delivery. Each gap can be closed by extension or by separate policy, but only if identified at placement. The largest uninsured losses on European construction sites typically fall into the materials-theft and delay categories, and operators frequently discover the gap only at the claims stage when the wording is read in detail for the first time.

How do brokers compare them?

Brokers typically produce a matrix of perils against policies, with ticks and crosses indicating coverage. This format is useful as a starting point and misleading as a conclusion, because a tick in one column does not have the same meaning as a tick in another. A serious comparison runs at the level of specific claims scenarios, with reference to the proposed wording, the deductible structure, the warranty conditions, and the case law interpreting the form. Operators should require this depth of analysis at placement and treat the matrix as the agenda for the conversation, not the answer.