"More Cowbell"

At the end of January, Bill McBride (Calculated Risk) picked up some of the data economist Tom Lawler (Lawler Economic and Housing Consulting, LLC) reported from the 10Qs of selected publicly-traded builders (Calculated Risk: Lawler: Some Home Builder Results) highlighting the trends in net new orders and backlog.

The definition of those terms in the SEC reports (net orders is new orders minus cancellations, order backlog is number of homes ordered, but not delivered) raised my curiosity about how that reporting impacted my area of interest, which is the operating performance and production management practices of homebuilding companies.

After all, there is an operational distinction in the homebuilding industry between delivered and ordered, and orders are not counted as work-in-process;  orders are a backlog of job starts.  And, if work-in-process is being reported as part of public builders’ backlog, it would represent a significant disconnect between the way they report their financial information and the principles that govern production.  The amount of work-in-process a building company carries is a big deal;  it has a lot to do with cycle time (because process duration is a function of the relationship between work-in-process and throughput/output), and has, therefore, an enormous impact on ROA.

So – I decided to take a look at D.R. Horton’s 10Q for 2012 Q4.  The terms “net sales orders”, “sales order backlog”, and what Horton terms “homes in inventory” have distinct meanings, but the first two terms do not create the disconnect between financial reporting and production physics that I suspected there would be.

But, there is this, regarding work-in-process:

Horton reportedly had 14,200 units of “homes-in-process”, 2,400 of which were completed homes.  Take those completed homes out, and the net WIP of 11,800 units against quarterly settlements/closings of 5,182 units indicates a construction cycle time of 205 days.  The reciprocal of cycle time is inventory turn, and based on annualized closings at that level (and it was likely lower than that earlier in 2012), inventory turn would have been about 1.75, close to what would be expected from a 205 day cycle time.

I would have expected Horton to be generating that level of closings on 7,000 units of work-in-process, probably less;  I would have expected their cycle time to be shorter;  I would have expected their inventory turns to be faster.

As for the impact of work-in-process on ROA, the expanded version of the formula for calculating Return on Assets is Return on Sales x Asset Turn;  according to DuPont, ROA is margin x velocity.  Certainly, not every asset on a balance sheet is inventory-related and affected by asset turnover, but the majority of Inventory in a homebuilding company tends to be its work-in-process (Horton is actually an exception, holding only 38% of its inventory-related assets in work-in-process;  NVR holds almost 90%).

As an industry group, I wonder if homebuilders pause to sufficiently consider the impact velocity has on economic return.

That is the point some of us have been making to this industry.

I wondered how Horton’s 205 day cycle time calculation compared with other public builders.  I looked at Pulte Group’s 10Q for the quarter ending 09-30-12.  Pulte reportedly had 7,927 units of “homes in production”, similarly defined to Horton’s “homes in process”, of which 566 were completed homes.  Applying the same method used with Horton, that would make Pulte’s net WIP 7,361 units, against quarterly settlements/closings of 4,418, which would indicate a construction cycle time of 150 days.

As it happens, Pulte’s inventory turn is 2.4, compared with Horton’s 1.75.

I also looked at NVR’s 10Q for the quarter ending 09-30-12.  Interestingly, NVR does not mention the number of units under construction anywhere in the management discussion of their SEC reporting, so you can’t look at it the same way.  But – given the success of the NVR model – I’m betting that the calculation of cycle time and inventory turn would have been better than either Pulte or Horton.

And, what does all of this say about productivity?

Productivity is the relationship between what is produced and what has to be consumed in order to produce it.  From any managerial standpoint, from any industry standpoint, from any enterprise standpoint, from any expert or business leader standpoint, the conventional, accepted formula for calculating productivity is Revenue divided by Operating Expense.  Less commonly, you will also find productivity expressed as the ratio between the input and output of a process.  Either way, productivity is about what is produced and what is consumed.

Unlike cycle time and inventory turn, the way productivity is measured has more of a financial connotation than an operating connotation.  The “what is produced” part is clear;  we understand what is meant by “output”.  What about “input”?  What is it that is consumed?  Is it an asset, or is it a resource?  Assets are converted or transformed.  Resources are consumed.  Inputs are expenses, which agrees, conveniently, with the first formula.

On that account, Horton had a productivity ratio of less than 9:1, Pulte’s was at slightly more than 10:1, and NVR’s was above 11:1.  Horton is bigger, but NVR is significantly (22%) more productive.  All three operating measures – duration, inventory turn, productivity – are connected.  The trend of one usually indicates the trend of the others.   Apply those differences in productivity ratios across billions in Revenue, and pretty soon you’re talking about serious money.

Which gets us to two final points:

First of all, D.R. Horton, Pulte Group, and NVR are not the issue.  They are three publicly-traded builders that all have to report their numbers;  the point isn’t whether they are good or bad, or better or worse than each other or other building companies.  The numbers illustrate the operating performance of real companies in the real world, and they enable us to learn from what we see.

For the past four years, the consistent message of Escape from Averageness, to builders, has been this, simply:

Do not become just another building company trapped in tar-pits of averageness, satisfied with having adopted the “best practices” of other builders, content to be good, no-better-but-no-worse than the other builders with whom you compete, a building company with a middle-of-the-road approach to delivering the value your homebuyers demand.

In the end, building companies should not want to be judged by what they are or what they accomplish;  they should not want to be judged by what they are or what they accomplish, compared to other building companies.  They should only want to be judged for what they are and what they accomplish, compared to what they could have been and could have accomplished.

The second point is this:  There is a perception about the advantage that rests with the size of balance sheets, particularly with cash and inventory;  and, yes, the capacity to borrow and the cost to borrow.  Size enables, but size also tends to foster a “more-with-more” mentality;  large homebuilding companies rarely exploit the advantages of becoming more productive, of embracing a “more-with-less” mental model.

The weapon of choice becomes more of what they have already.  More geography;  more communities;  more inventory;  more cash;  more debt;  more overhead.

“More cowbell.”

The legendary Bruce Dickenson – yes, the Bruce Dickenson – would encourage us to “explore the space.”

The clearest image we have of homebuilding production – the best visual reference – is that of a pipeline, one in which size is defined by the amount of work-in-process the pipeline is designed to carry, in which cost is determined by what the production effort consumes, in which length is calculated as the pipeline’s cycle time, and in which capacity is defined as the rate of throughput a pipeline of that size can produce, with a planned, finite, and controlled amount of work-in-process.

The amount of Inventory a building company has to carry in order to generate its Revenue is the more meaningful, more useful measure of its true size.  Using work-in-process as the measure of size would discourage builders from aiming to be bigger companies;  it would convince them they need to become faster, more productive companies.