Defence Contractors Have One Customer and Very Long Memories

The Customer Concentration

A prime defense contractor derives the vast majority of its revenue from a government plus export sales to allies that the government itself has yet to approve. The first time I opened a Prime's files that single fact explained the whole thing more than anything else on the page. Almost all of the strange things in the behavior of these companies can be traced back to this

In almost any other industry relying on a single buyer for the majority of your revenue would be considered an imminent risk. It's simply the way of the market here and it drives behavior not seen anywhere else. Contractors put real effort into the budgeting process itself. They spread facilities and suppliers across a broad map of political districts. They treat the customer relationship as a core capability staffed and managed for decades rather than as a sales team that has to close deals every quarter

My read is that the political map matters almost as much as the technology. A program with work spread across enough districts becomes something a legislature doesn't want to cancel because canceling it involves a local cost to specific representatives who then have to explain the jobs lost in their countries. That's a customer-driven strategy not a product-driven one and once you notice it you start to see it across the sector. This is also why a program that's behind schedule and over budget can still be very difficult to kill

Contract Types Determine the Risk

Cost-plus contracts reimburse allowable costs and add an additional fee. Governments use them when the work is truly uncertain simultaneously limiting the contractor's disadvantages and advantages. Firm fixed price is at the other extreme. The contractor sets a price and if the work is overrun that excess comes directly out of his own margin. Fixed-price incentive contracts are in the middle sharing the overages up to a maximum limit above which the contractor is alone

Governments have spent years pushing work toward fixed prices to control their own spending and on paper that looks like pure discipline. The part I think is underrated is the failure mode. A contractor that offers a fixed price for development work that it hasn't fully covered is effectively writing an option for the government and that option becomes very expensive when the engineering turns out to be more difficult than was assumed in the bid. The useful habit here is simple. Read what type of contract a major program is run on and you'll immediately knowWhere the excess risk lies. It is revealed so you don't need to guess

Backlog Is the Real Metric

Because programs run for years or decades a single quarter of revenue means almost nothing. The numbers that really tell are the total backlog and within that the split between funded and unfunded

Funded order book He has appropriated the money that already supports him. Unfunded delay It is a contracted amount that still depends on future budget allocations that no one has approved yet. They are both real and not of the same asset quality. A prime company can post a weak quarter and still be in a strong position if the backlog is deep and it is heavily funded. It can post a good quarter and be quietly weaker if the backlog is narrowing or if it relies on allocations that have yet to survive a political fight

A Worked Example: Reading the Backlog

Let me nail this down with round illustrative numbers rather than a single actual company. Let's say Prime reports $40 billion in annual revenue and $60 billion in total portfolio. The first thing I calculate is coverage: 60 divided by 40 is 1.5 so the lag represents about a year and a half of revenue that's already under contract. That's the visibility the market is really paying for

Next I look at the book to bill which is the new booked orders divided by the revenue billed in the same period. If the company booked $45 billion in new orders against that $40 billion in revenue the book value to bill is 45 divided by 40 or about 1.13. Anything above 1.0 means that the backlog is growing faster than it is being consumed so the pipeline is filling up notdraining.Below 1.0 the company lives off the work it has already earned

Then I divided up the backlog. Let's assume that $38 billion of the $60 is funded and $22 billion is unfunded. That's about 63 percent of the funding meaning more than a third of contracted work still relies on credit that hasn't been secured. In a calm budget environment that unfunded third is close to money in the bank. In a year of budget fights or continuing resolution it's exactly the pieceWhich I would test first because it's the part that a political process can quietly eliminate. Two companies can show the same headline of 60 billion and be in really different shape once this split is made

Case Study: Boeing and the KC-46

The clearest lesson I've ever learned about contract type is Boeing's KC-46 Pegasus the refueling tanker built on a fixed-price development contract. Boeing won the award in 2011 with an aggressive bid betting that it would be manageable to retrofit a commercial 767 into a tanker. The engineering especially the remote viewing system used to guide the refueling boom turned out to be much moredifficult than the offer entailed

Because the contract had a fixed price the government's exposure was limited and Boeing absorbed the cost overruns. By the early 2020s Boeing had collected billions of dollars in accrued pre-tax charges for the program a figure that continued to rise as new technical problems arose. That's the way fixed-price fails in a program: the same structure that looks like taxpayer discipline turned a big win into years of losses for the contractor. The customer got its cost cap. The shareholderpaid for optimism in the offer

What I gather from the KC-46 is not that Boeing is exceptionally bad at this. It's that the type of contract is a genuine predictor. When I see a company with the highest bidder pricing a job that is still in development I now treat it as a red flag about future margins not a sign of discipline

The Barriers to Entry

Technical capacity matters and it is not the wall that keeps competitors away

The real barriers are the ones that take years to remove even after you can build it. Security clearances. Facility accreditations. Compliance with government cost accounting standards. A track record of delivered programs which by definition you can't have until you've already been entrusted with the programs. A well-funded startup can hire engineers. You can't overnight conjure up a decade of licensed facilities and audited accounting history and that's the part that really protects incumbents

Consolidation has left a short list of top candidates capable of leading the most important programs. That gives them bargaining power and at the same time worries the customer because the government has a real interest in keeping more than one capable supplier alive. That tension is why a defense merger is sometimes blocked or a second supplier financed at a cost that a single source would not bear. In practice the customer is paying to preserve the competence it might need later

Why Margins Are Moderate

Defense margins tend to be lower than you would expect for companies with such strong competitive positions. That surprised me at first and the explanation is simple once you see it

The customer knows his costs. Government contracting comes with audit rights cost accounting standards and pricing rules created specifically to prevent excessive profits on work that is not bid competitively. When the buyer can read your books and has written the rules about allowable profits he will not be able to set the price like a company with a moat even when it clearly has one

What you get in return is predictability long duration and almost no credit risk for the customer since the counterparty is a national government. In my view a defense premium is closer to a bond than a high-yield stock. You are buying long reliable flow rather than explosive growth and the moderate spread is simply the price of that reliability

The Cycle

Defense spending follows geopolitical conditions but with a delay because budgets are set annually and programs take years to increase or decrease

Delay comes in both directions and I think that's the part that trips people up. A jump in threat perception manifests first as budget growth and only several years later as contractor revenue once the programs are funded staffed and produced. A stretch of reduced stress produces outages that take the same time to pass through the same pipeline. So a contractor can position itself perfectly for the latter environment and spend years repositioning itself for the current one. When you read this sector actuallyYou read credits from a few years ago not today's headlines

Where This Model Breaks

I've made the bullish case for defense premiums as durable bond-like businesses so let me honestly argue against them because the risks are real and specific

The first is program cancellation. The political difficulty of killing a program is a trend not a law. The Army's Comanche helicopter and Crusader artillery system were canceled after years of spending and Future Combat Systems was largely dismantled. When a marquee program dies the backlog behind it can evaporate faster than the models assumed. The second risk is the fixed-price trap of the KC-46 story expanded. A premium that has bid several programsof fixed price development at a time can turn a supposedly safe business into a sequence of charges. The third is the budget risk at the top. Sequestrations and repeated ongoing resolutions have shown that a single client can freeze or cut expenses for reasons that have nothing to do with the performance of any contractor

The last one is the most interesting to me. The barriers to entry are high but not eternal. New entrants in launch software and autonomous systems have begun to clear the licensing and accreditation hurdles faster than the traditional playbook assumes and are competing on cost structures that the incumbents can't easily match. If I'm wrong about defense cousins ​​being safe this is probably where the error lies: assuming that a moat stands simply because it has stood for forty years

How I Actually Analyse a Prime

If I was given a defense department's annual report and a time this is the order I would work in because it's very different from how a normal industrial company would read it

I would go to the backlog table before the income statement. The total backlog the funded and unfunded split and the book to bill tell me more about the next five years than this quarter's revenue. I would then find the contract mix cost plus versus fixed price and pay special attention to any major fixed-price development work as that is where the unpleasant surprises are born. Only then would I look at the margins and interpret a moderate margin as a characteristic of the audited and regulated relationshiprather than a weakness. Lastly I would like to assign the most important programs to their political exposure and ask what allocations each depends on. My honest opinion is that these are companies that are funded based on the durability of the backlog and the politics behind it and the revenue line that everyone quotes first is close to the last number I would rely on

The Bottom Line

Defense contracting depends on a dominant customer that sets terms audits costs and keeps margins low and pays for that leverage with programs that last a decade and with almost no credit risk. Contract type tells you who eats up the excess and the KC-46 shows how costly it can be to get it wrong. The order book matters far more than any quarter the division between funded and unfunded is where political contingency lives and a close look at coverage and billing will tell you more than topline revenue.model is durable but not foolproof: cancellations fixed price traps budget cuts and a new generation of entrants are all real. My conclusion is that these companies are judged by the durability of the backlog and the policies that support it not by the top line that everyone cites first