As more dwellings are approved for construction, average house completion times typically push-out. That appears to be holding true over recent times, but why is it so? And what are the implications of longer ‘approval to completion’ cycles, especially when we complete the same number of houses but take longer to build them? These were some of the questions explored at a recent workshop convened by FWPA, in Canberra, in conjunction with ABARES.
This is a question of ‘known unknowns’, one might say, because many participants in the housing market need to understand cycles at an ever more refined level, and thus have many questions (or the same questions cast a different way, at least) about housing cycles, when they will turn and what it means for them and their business.
As FWPA’s Economics and Statistics Manager, Jim Houghton advised at the outset of the workshop:
“A key driver of demand for sawntimber products is residential housing construction. However, housing construction is a cyclical business. So the holy grail in our industry is trying to understand what indicators may exist to identify these turning points.”
To accentuate this point, the following chart was provided.
It was agreed that historically, many of the housing cycles have been linked to interest rate cycles, especially as the economy reaches its full capacity.
However, that does little to explain why at different points in time, the average time it takes to build a detached house rises so significantly.
This has been the subject of further work by Dr Ehsan Gharie, RMIT who has been a member of a team looking at this issue from the perspective of housing affordability. For the timber industry shorter completion times assuming demand is steady would represent opportunities for further sales. This research was presented at the workshop and provides possible insight to the reasons.
As we can see in the graph below the completion time to build a house has increased from 5 months in 2002 to 7.5 months in 2009 and eased slightly 6.5 in December 2017.
The logical answer to this increase in completion times would be we are building more houses. However that is not the case. We are in fact completing about the same number of houses but it is taking longer. This can be seen in the chart below, which compares the number of house completions and the average house completion time for the same period.
The chart shows that the output of the industry remained stable over the period. The house building industry produced on average about 26,000 houses per quarter between 2001 and 2017. However, during that time, completion times increased.
Dr Gharaie commented:
“In other words, the 26,000 houses in 2001 were built on average in 5 months, while the same number of houses in 2009 were built in 7.5 months.”
This has eased to 6.5 months by December 2017 but is still some 30% above the rate in 2001.
Another common explanation for increased housing completion times is that housing sizes increased. The data indicates there was an increase in house floor areas, but that it does not match the increase in completion time, as the chart below shows.
So, these factors do not impact housing completion times.
Dr Gharaie subsequently used manufacturing analysis to consider housing-under-construction from a ‘work in progress’ perspective, where there is a correlation between work in process (WIP) and cycle time (CT) in a production line. WIP in house building industry is measured by ‘number of houses under construction’ and CT is measured by ‘average house completion time’.
Dr Gharaie wrote:
“…if we perceive the house building industry as a production system, we should be able to see a correlation between number of houses under construction and average house completion time.”
In manufacturing systems this relationship is informed by Little’s Law which shows the interconnection between these factors and in theory the cycle time (completion time) increases when you move beyond Work in Progress (houses under construction) capacity.
Dr Gharaie has then applied this principle to the Australian house building industry and confirmed the same “two phase” relationship exists.
Dr Gharie, commented,
“The break point or the critical number of houses under construction falls on 48,000 houses. This is a rough figure; however, it points out to an important concept. It shows the industry capacity is limited. Whenever, the number of houses under construction goes beyond this level the completion time increases and whenever the number of houses under construction is below this figure the completion time is at its minimum level.”
In addition Dr Gharie’s research indicates that when operating at capacity then the min housing completion time is 5.25 months (1.75 quarters).
The chart below shows this relationship.
There is a clear correlation between the number of houses under construction and house completion time. This strongly suggests that the housing sector works like a production system. When the number of houses under construction (WIP) rises beyond a “steady state” level of capacity, average house completion times increased. Or another way of looking at this is as demand (houses under construction) increases beyond current capacity the same number of houses are being built (completed) but it is taking more time to do it.
Or another way of looking at this is if you put more work into the system (houses under construction) in circumstances where the sector is operating at capacity then you will only get the same number of units out the other end (housing completions). However, what has changed is the time taken to build (completion time) is longer. This is a big challenge as the longer completion time will usually be associated with higher costs – the old analogy time is money.
This is a very useful start to understanding capacity in the Australian housing sector, and its relationship to construction time. It does not address at this time, the questions of why exactly this is the situation, but it is information that sets the scene for further development work, as well as for some soul searching inside businesses.
What are the implications of a slower production rate at the top of the cycle, when capacity has been exceeded, or when ‘business is booming’ so to speak? Here are a few.
First, less homes can be built over a given time, meaning less new dwellings can be sold and lower total revenues.
Second, a slow house production system provides space for a supplier to disrupt the normal patterns of dwelling production, through an improved (faster and more efficient) house manufacturing model. Perhaps by an ‘installed solution’ or ‘offsite manufacturing’ option, they could remove the ‘friction’ that causes compounding of lost time when multiple trades are held up by one another. This manufacturing system of work concept may also provide an insight into the emergence of steel framing as a substitute product.
Third, and related to the second, costs associated with the ‘friction’ are likely to be reducing margins, costing more and making scarce skilled labour resources less efficient than a robust building cycle can accommodate.
This might be one of those known unknown situations, but it is not just some theoretically interesting question. The answers to housing cycle turning points, their drivers and indicators is vitally important to the future of the Australian housing economy.
