Virus-driven behaviour changes that support a faster online transition will boost industrial demand over the coming years. But we don’t believe the view that higher online spending will cause rents to detach permanently from the underlying strength of the economy. And since we don’t expect to see the sort of supply shortages that plagued the UK market in recent years, and supply capacity is even greater in the euro-zone and the US, any initial post-COVID boost to rents is only expected to be temporary.
- Virus-driven behaviour changes that support a faster online transition will boost industrial demand over the coming years. But we don’t believe the view that higher online spending will cause rents to detach permanently from the underlying strength of the economy. And since we don’t expect to see the sort of supply shortages that plagued the UK market in recent years, and supply capacity is even greater in the euro-zone and the US, any initial post-COVID boost to rents is only expected to be temporary.
- The structural change towards online spending was already benefiting the industrial sector before the pandemic. And it seems likely that new living and working arrangements after COVID-19 will further boost online spending and delivery needs. Our estimates suggest that these virus-driven changes in behaviour could increase industrial demand by an additional 15% in the UK and the US over the next decade.
- We think that higher online spending will provide more additional demand than virus-driven supply chain changes that also boost the need for industrial space. That is not to say that supply chains will not adapt. But we think it is more likely that these changes will be driven by factors such as technology and politics, which were already causing globalisation to falter before the pandemic, rather than the virus alone.
- This faster online transition is likely to have a more material additional impact on UK industrial demand compared to our pre-virus expectations than in the US. This is because, even before the virus outbreak, we expected US demand to increase substantially as online shares catch up. In contrast, industrial space requirements in the UK were expected to increase by proportionally less compared to the previous decade.
- Our estimates suggest that this increase in demand could be enough to push UK industrial rental growth towards 4% p.a. later in our five-year forecast horizon, compared to our pre-virus forecast of 2% p.a.. That said, as we have argued before, there is little evidence that higher online spending alone has led to a structural change in the pace of industrial rental growth in the past. And, given the supply backdrop in the UK is less constrained than in recent years, we don’t think that rents will grow permanently at this higher pace over the next decade. The exception could be industrial property near population centres, where demand for delivery is greater, costs are higher, development has been lower and land is more scarce.
- In aggregate, we also expect that supply will be able to adjust quickly in the euro-zone and the US. Indeed, in the euro-zone, the ability to service demand through cross-border deliveries means supply potential is greater than the UK. And even in Germany, there have been no signs that prime industrial rents in the key markets have detached from the pace of regional GDP growth, even though the online transition has been similar to the UK. That said, in some US markets, notably Chicago, LA, New York and the Inland Empire, industrial rental growth has outpaced the strength of the economy in recent years. These would be candidates for where to look to for signs of pressure on industrial rents over the next decade.
- This is the fourth publication in our Future of Property series, considering the long-term outlook for real estate. If you would like to learn more about the series, please contact your account manager or sales executive.
Higher online share to boost logistics demand by 15%
While the virus outbreak has brought into question the future of the office and has spelt disaster for an already struggling retail sector, the general view is that it bodes well for the industrial sector. This reflects that it has encouraged a shift towards online spending, which, if sustained, could materially increase demand for industrial space.
In this Focus, we explore what virus-driven changes in behaviour towards online spending might mean for industrial demand over the next decade and what these could mean for rents. To this end, we also consider the outlook for supply and how it could influence rental dynamics.
As in our previous Future of Property publications, we assume that the initial threat of COVID-19 will be removed. This means our estimates are based on the view that the rise in short-term requirements driven by the virus outbreak and the immediate impact from the reduction in economic activity subside.
We focus predominately on the UK and the US. However, where relevant, we draw on the experiences of other major commercial property markets. We also concentrate on the impact on warehouse and logistics property, as it the most important investment asset and, in our view, is most exposed to any behaviour changes caused by the virus.
All roads lead to the shed
The starting point for this analysis is the conclusions from the previous publications in our Future of Property series, which provide useful insight into how much additional industrial demand could be generated.
The first is that the virus outbreak will encourage a faster online transition. As discussed in our Future of Property Retail Focus, there is a strong argument that the virus will encourage a quicker adoption of online channels, particularly for categories where the online share had previously been low, but which have seen substantial growth this year, such as groceries, and health and beauty products.
What’s more, changes in the way we live and work after COVID-19 could also allow online penetration to reach a higher level. The conclusions from our Future of Property Office Focus suggest that there will be a material shift towards home working. Although we argued in our Future of Property Residential Focus that this may not lead to a mass exodus from living in the cities in the longer term, it is likely to increase willingness to shop locally or order online. Indeed, instead of picking up items at the office, home delivery is generally more convenient for a wider range of items. This is especially the case for products like groceries, which, given their perishable nature, are practically difficult to deliver unless someone is home to receive them.
These will be considered against other potential changes to demand for space. In particular, the virus outbreak has highlighted the importance of supply chain resilience, prompting calls for holding more inventory or reshoring or nearshoring production.
The faster online transition and increases in supply-chain resilience would point to greater demand for industrial space. The question is to what extent will significant changes be made? And by how much could industrial demand increase as a result?
In the near term, more resilient but not unscathed
Demand for industrial property has not been immune to the virus outbreak, given the resulting supply chain interruptions and the contraction in economic activity. But it has been more resilient than in other property sectors. However, across key markets there have been stark differences in the impact on occupier demand, even within the same country. (See Chart 1.)
Chart 1: Industrial Take-Up (2020 H1, % y/y)
Sources: Colliers International, REIS *Net Absorption
The UK provides a good example of this divergence. In aggregate, industrial demand held up well over the first half of the year, boosted by short-term requirements from the likes of Amazon, supermarkets and the National Health Service. But industrial property around London has suffered. This reflects that occupiers in this market have been more exposed to the hard-hit retail, hospitality and leisure sectors.
That said, once the virus is brought under control, our view that there will be a faster online transition means that this market should benefit from a boost in demand for last-mile delivery. This highlights the importance of taking a longer-term view. We now consider what this could look like in more detail.
Longer term, how could the virus change demand?
In what follows, we estimate how much additional industrial space could be needed over the next decade because of the behaviour changes caused by COVID-19, focusing on the UK and the US.
Here, we concentrate on demand for space from retailers and the food industry. This is because we think they are most likely to see online shares shift higher because of the virus, particularly in the grocery and health and beauty industries. In addition, these two sectors combined account for the largest share of logistics take-up. (See Chart 2.)
Chart 2: UK Logistics Take-Up by Sector (2019, %)
To proxy demand for industrial space from retailers and the food industry, we use our macroeconomic forecasts for online and offline retail sales. With this we estimate how industrial demand might change as online demand increases, while accounting for any displaced demand from lower spending instore.
We combine these forecasts with Prologis estimates that, for each additional $1bn of online retail sales in the US, retailers need another 1.2m sq. ft. of industrial space, while an additional $1bn of offline retail sales would be consistent with a smaller figure of an extra 425,000 sq. ft. of industrial space.
For the UK, estimates differ across the various agents. For example, every £1bn of online retail is estimated to result in the addition of around 900,000 sq. ft. of industrial space by CBRE and more than 1.3m sq. ft. by Knight Frank. For our purposes we used an estimate in the middle of these.
This rule of thumb that industrial space requirements are three times higher for online reflects that retailers that sell online tend to offer more product variety and need more stock buffers. Further, a greater amount of space is needed as they generally ship directly to the customer and allow for returns.
Admittedly, this is a simplification and across individual retailers and markets there will be variation in this figure. It will depend on the type of retailer and where they are in their development cycle. But we think it is a reasonable figure to use in aggregate and it is likely to remain valid.
Indeed, although some will reach maturity, new forms of demand will arise. For example, if we are correct that online grocery shopping will receive a boost, this could increase space requirements. This is because higher volumes would allow what is currently a labour-intensive process of picking items instore to shift to dedicated distribution centres and smaller warehouses near population centres, possibly like the Ocado model.
To make an assessment of how much additional space will be needed, we first need to understand what space would have been needed if the virus didn’t cause any behaviour changes, or in other words if online penetration continued its pre-virus trend – this is our baseline scenario.
Before the virus outbreak, we assumed that the online share of retail sales would gradually increase from its 2019 share of 19% in the UK to 30% by 2030, and from 11% to 25% in the US over the same period. (Note that given US retail sales also include spending on vehicle fuel and repair, these online shares are not directly comparable.)
Our estimates suggest that this would mean that around 100m sq. ft. and 1,800m sq. ft. of additional industrial space would have been required in the UK and US respectively over the next decade. (See Chart 3.) As a share of stock, the increase is much larger in the US, reflecting a stronger outlook for retail sales growth and that online shares are catching up to the UK.
Chart 3: Baseline Additional UK and US Industrial Space Required (M Sq. Ft.)
Source: Capital Economics
In the UK, the outlook for the next decade is only slightly stronger than our estimates of the additional space that was required in the previous decade. The main difference is that demand for industrial space from bricks and mortar retailers is expected to be lower as offline retail sales are forecast to grow at a slower pace.
In the US, given the lower starting point for online penetration, the increase in required industrial space is much larger compared to the previous decade. In fact, around 80% of our estimate of required space is attributable to online sales, compared to about 50% in the previous decade.
A similar exercise for the big four euro-zone markets shows that our pre-virus expectations for online penetration and our macroeconomic forecasts for retail sales would imply slightly more industrial space would be required in Germany than in the UK over the next decade. (See Chart 4.)
Chart 4: Baseline Additional Euro-Zone Industrial Space Required (2020-2030, M Sq. Ft.)
Source: Capital Economics
However, given our view that structural factors, such as preferences and technology, would prevent online penetration taking off in southern Europe, the required space directly linked to retailers in Spain and Italy is expected to be lower.
Of course, these estimates of required space may not end up being the amount of space actually taken. But history suggests that our figures are plausible. Indeed, in the UK, Savills estimate that between 2009 and 2019 logistics take-up totalled around 300m sq. ft.. Given that retailers have accounted for about 30%-40% of take-up over time, our estimate of required space over this period seems reasonable.
We use the conclusions from our previous Future of Property work to construct a scenario for the outlook for industrial demand in the UK and the US, which incorporates our view of virus-driven behaviour changes. This scenario assumes that the virus outbreak brings forward about two years of ‘normal’ online growth relative to our baseline, but also that changes in the way we work and live allow online penetration to reach a higher level. We assume this ‘new’ level is around 35% in the UK and 30% in the US.
The results suggest that that demand for industrial space post-COVID could be about 15m sq. ft. and 250m sq. ft. higher than our baseline in the UK and the US respectively over the next decade. (See Chart 5.) In both cases this is a boost to industrial demand of 15% relative to the baseline.
Chart 5: Baseline Versus Scenario Industrial Space Required (M Sq. Ft.)
Source: Capital Economics
One key point to note is that this 15% additional rise in space required is a less significant change in the US, as pre-virus we would have already expected industrial demand in the next decade to be 80% higher than in the last decade as the online share catches up with the UK. In comparison, our pre-virus estimate of required space in the UK over the next decade was for an increase of just 15% compared to the 2009-2019 period. As such, the relative increase in virus-driven demand for space will benefit the UK proportionally more.
A similar exercise is more complicated for European markets given the differences in online development between countries, as well as data constraints. However, we highlighted that a faster online transition is more likely in the Benelux, German and Nordic markets where online penetration is at similar levels to the UK. Indeed, early evidence suggests that this could be the case in Germany. (See our Update.)
But at the same time, it is less obvious that these markets will experience significant changes in living and working arrangements. Sweden, Denmark and Benelux already have a high share of remote working in office-type jobs, so a substantial shift here appears less likely. That said, the low share of remote working in Germany and Norway suggests that these markets could be candidates to watch for a larger post-COVID boost to industrial demand. (See our Focus and Update.)
What of pandemic-driven changes to supply chains?
The pandemic has also highlighted the importance of risk management in the supply chain process. In turn, many have argued that it could have longer term impacts in the form of firms holding more inventory, reshoring or nearshoring, particularly for critical or cost sensitive goods.
Estimates suggest that the impact on industrial demand of holding more inventory alone could be quite large. Indeed, Prologis suggests that inventories could be 5% to 10% higher under a ‘just in case’ inventory model in the US. Even holding sales constant, this could result in an additional 285m-570m sq. ft. increase in space requirements.
This is not insignificant; the upper end of this range would be equivalent to 30% of the industrial space we calculated would be required by retailers in the US over the next decade. That said, the impact would have been even greater had there not been the end of the downward trend in the inventory to sales ratio following the Global Financial Crisis. (See Chart 6.)
Chart 6: Retail Inventory to Sales Ratio
Sources: US Census Bureau, Refinitiv, Capital Economics
Although some near-term increase in inventory is likely until the virus is brought under control, we think it is unlikely that there will be significant changes to supply chains longer term because of the virus alone.
Indeed, there are cost pressures and practical constraints that need to be considered. For one, the dominance of China in trade and the complexities of modern supply chains means that even if there was a shift, it could not happen overnight. Further, there are capacity constraints to moving production back onshore on a large scale.
This is particularly the case in Europe, where the modern warehousing stock per capita is low. (See Chart 7.) That said, within continental Europe there is a greater ability to service online customers via cross-border deliveries than in the UK, which boosts supply potential for multinationals.
Chart 7: Modern Warehousing Stock Per Capita
But even though there is more capacity in the US, it is not necessarily the case that reshoring will increase. Indeed, broader measures of manufacturing costs, which along with the physical cost of production include productivity growth, currency exchange rates and energy expenses, show that there is still a significant advantage to manufacturing outside of the US. (See Chart 8.)
Chart 8: BCG Global Manufacturing Competitiveness Index (2019, US=100)*
Source: BCG *Below 100 = More competitive than the US
What’s more, reshoring indices show that it was not until 2019’s significant change in US trade policy that there was any material shift away from manufacturing in China. But even this drop was made up for by imports from low-cost Asian countries and Mexico, rather than any increase in domestic manufacturing output.
And most importantly, past experience suggests that these shifts are not likely to be driven by the pandemic alone. In fact, our Focus argued that previous supply chain shocks, such as the 2011 Japanese earthquake and Thai floods, did not lead to significant reshoring or reorganisation.
Overall, this does not mean that there won’t be significant changes in supply chains. However, these are just as likely to be driven by other factors such as changes in technology and politics, which were already causing globalisation to falter before the pandemic.
How could this impact industrial rents?
On balance, we therefore think that the more material increase in industrial demand because of virus-driven changes will come from the shift to more online spending. To understand what this might mean for industrial rents we focus on the UK, since this is where we think the additional benefit will be greatest.
Our estimate of industrial space required by retailers in the UK has had a reasonable relationship with industrial rental growth in recent years. (See Chart 9.)
Chart 9: UK Retailer Demand for Industrial Space and All-Industrial Rents (Annual)
Sources: MSCI, Capital Economics
Admittedly, this estimate has been volatile. But the relationship also held in recent years when growth in industrial rents detached from the underlying pace of economic activity. (See Chart 10.)
Chart 10: UK Industrial Rents and GDP Growth
Sources: MSCI, Refinitiv
At face value it suggests that, following some volatility during 2020 to 2022 as the level of online penetration normalises, virus-driven changes in behaviour pose upside risk to our five-year forecast for industrial rental growth in the UK. Implied rental growth of up to 4% p.a. would be more than our current projections. (See our UK Commercial Property Outlook.)
That said, this measure does not account for the impact of supply. We have argued previously that it was supply shortages, rather than rising online penetration and the associated increase in occupier demand, that had for the most part caused the divergence between industrial rental growth and UK economic growth in recent years. (See our Focus.)
And we don’t expect to see a return of the sort of supply shortages experienced between 2011 and 2016. This is because developers in the UK had already responded in the pre-COVID period and rental growth had slowed to more sustainable rates after 2018. (See Chart 11.)
Chart 11: UK Industrial Development Completions
(M Sq. Ft.)
Source: Gerald Eve
And we expect that strong industrial performance will continue to encourage development post-pandemic, even without strong rental growth, given the headwinds in the retail and office markets. As such, although the increase in industrial demand will provide some support to rents, we expect that the boost will be temporary as supply responds.
While overall supply may be responsive, one area where supply could be more restrictive is last mile. Given the value placed by consumers on rapid delivery, we expect there will be more upward pressure on rents near large population centres, particularly in London and the South East. These areas are where demand for delivery is greater, costs are higher and land is scarce. Indeed, there has been a strong relationship between land values and industrial rents in the UK. (See Chart 12.) What’s more, in general there has been less development of these smaller units, so supply is expected to be more restricted.
Chart 12: UK Industrial Rents and Land Values
Source: Carter Jonas
Of course, some will argue that conversions of property types that are in excess supply, such as retail, could alleviate this pressure. However, there has historically been a wide gap between the suitability of retail and logistics premises. Meanwhile, residential demand is not expected to falter.
Admittedly, given the downward trend in retail capital values in recent years, at least on an economic basis conversions of cheaper retail warehouses or mixed-use approaches are starting to look more viable, particularly in London and the East, where industrial rental values are higher. (See Chart 13.)
Chart 13: UK Retail Warehouse Capital Values Relative to Industrial (2020 Q3)
But there are also practical, political and legal constraints to consider. This led Prologis to conclude that retail-to-logistics conversions will be low and time frames long. In fact, across their US markets, they expect conversions to account for less than 3% of annual completions over the next decade. In the UK, this figure could be a bit higher given land is more constrained. But even a result double this would be small.
Overall, at the aggregate level, it seems likely that industrial rental growth could be a bit stronger than expected later in our five-year forecast horizon, particularly near large population centres. But we think supply will prevent rents growing permanently at this higher pace over the next decade.
What could this mean for other markets?
Unlike the UK, there have been fewer signs in the US or euro-zone that industrial rental growth has detached from the underlying strength of the economy in the past. This suggests that supply has been able to keep up with occupier demand, even as online penetration has risen. This is consistent with our view that the US and continental Europe are less supply constrained.
Indeed, within the euro-area, the German markets, where online penetration is close to that in the UK, have not seen industrial rental growth detach from the strength of the economy. (See Chart 14.) In contrast, Barcelona industrial rents have taken off despite online penetration being low, suggesting that it is not only demand driving rents.
Chart 14: Cumulative Industrial Rental Values and Regional GDP Growth (%, 2014-2018)
Sources: Refinitiv, Eurostat, BEA, ONS, REIS, MSCI, CE *MSA
That said, in the US, it is notable that industrial rental growth in Chicago, LA, New York and the Inland Empire metropolitan statistical areas, have outpaced regional GDP growth in recent years. We don’t think that virus-driven demand will play a material role in driving rents, given the substantial increase in occupier demand already expected in absence of the pandemic. But these markets stand out as where pressure on industrial rents could be more likely over the next decade.
Bringing all of this together, retailer demand for industrial space in the UK and the US was expected to be greater over the next decade than the previous decade, even in absence of the pandemic. But our calculations suggest that the virus-driven behaviour changes that allow a faster online transition could boost logistics demand by an additional 15%.
The relative increase in virus-driven demand for space is expected to benefit the UK the most. This is because we already expected before the pandemic that industrial demand would increase substantially in the US over the next decade as the online share catches up with the UK.
The impact on rental growth in each market is likely to depend on the ability of supply to adjust. In the US and euro-zone, supply capacity appears greater than the UK, which suggests that in most markets supply will be more able to keep up with occupier demand, even as online penetration rises. But even in the UK we don’t expect to see the sort of supply shortages that plagued the market in the late 2010s. As such, we think that any initial post-COVID boost to industrial rents will be temporary.
That said, with much of the increase in space requirements expected to be driven by increasing home deliveries, much of this demand is likely to be near large population centres. Given that we don’t expect a material exodus from cities in the longer term, there is likely to be more significant pressure on rents here.