John Lewis
Increased working from home (WFH) for public health reasons during the pandemic has spawned a debate about whether this shift might become permanent. In this post, I try to sketch out some of the (macro) economics of a longer-run post-pandemic shift towards more WFH. I argue that: i) on consumption, it won’t affect aggregate expenditure, it will just reallocate it across space and sectors ii) in property markets, effects hinge on supply responses; iii) for output, cost-savings to firms from cutting back office space don’t translate one-for-one into GDP gains.
This post comes with three qualifiers. First, I don’t consider any public health or epi-macro type feedback effects of WFH: I focus on a post-covid world where these aren’t operating. Second, I make no predictions about *whether* a shift to WFH will happen – I just conduct a ‘what if’ thought experiment. Third, I don’t analyse whether WFH makes workers more or less productive at their jobs: the jury is still out on this, and it’s outside the realm of traditional macro, so I leave it for now.
Does greater WFH depress consumer spending?
No. The simplest theoretical baseline is that it changes the places or sectors where people spend money, but not how much they spend.
Many city centres have businesses which rely on footfall from passing office workers and could be (and have been) hard hit by a reduction in frequency/volume of office visits: such as coffee shops, lunch places or dry cleaners. But the claim that this will lead to lower aggregate consumer spending reflects a fallacy of composition.
Economic theory says consumers optimise in two respects: they work out their best path of spending over time, and then within each time period, they allocate that spending to maximise utility from what they buy. Shifts in preferences, or technological innovations that reduce spending somewhere, don’t destroy the spending, they just relocate it.
Some of that switching is purely spatial – home workers buy fewer coffees in city centres and instead buy more from outlets near their homes. Some of it is sectoral – money saved on dry cleaning or commuting to work, gets spent on other stuff. Of course the pandemic did see a sharp fall in consumption – but this was because consumers were physically constrained in the scope and timing of their purchases by lockdown measures which won’t be present post-pandemic.
Do cost savings on office space boost GDP?
Perhaps. But it’s less clear cut than first appears. Greater WFH, means a reduction in office occupancy. Assuming firms doing more WFH do cut back on office space then there is a cost saving accruing to firms. But improvements to a firm’s bottom line do not automatically translate into higher GDP for ‘UK PLC’ because the economics of national accounts is a bit more complex.
The effect on GDP depends on whether greater WFH changes the underlying ‘production function’ – ie how real estate, labour and capital are combined to produce output. And if so, how much extra output is created as a result of the ‘freed up’ office space.
First up, consider the case where the production function *doesn’t* change. To revisit a favourite analogy, suppose there’s an island where boat owners hire workers to go fishing. One day the owner tells staff that they need to provide their own nets. The owner sells off the firms’ nets, which are then rented or purchased by workers. Effectively the cost of providing nets has been shifted from employer to employee, but production technology is unchanged. If wages stay the same, what the owner gains in cost savings, the workers lose. Regardless of whether wages adjust, production hasn’t changed so output is unaffected.
In WFH terms, this is equivalent to the case where workers use the same amount of space (but have to acquire it themselves) and produce the same output. There’s no net saving of space economy-wide, no increase in output and so cost savings from WFH don’t boost GDP.
If however, those workers don’t need to acquire any extra space and can still produce the same output as before, then there *has* been a change in the aggregate production technology.
This might differ across workers: for people with a spare room functioning as a study which is otherwise unused in working hours, or those working elsewhere no extra space needs to be acquired. But for others – eg a young group of graduates sharing a flat – obtaining the extra space to work is likely to require economic resources being diverted towards that (as per the fishing analogy). And early estimates are that this could be sizable.
And an office doesn’t just consist of real estate, it is also the furniture, IT equipment and the heating/lighting. Even if space can be provided for at essentially zero marginal cost by workers, other elements cannot. Indeed some research has argued the lack of economies of scale in energy use in home vs office working can be substantial.
But let’s suppose not all of the cost savings to WFH firms are pure transfers. So in aggregate there is some now ‘spare’ office space freed up, which can be put to new uses by others. In national accounts terms, the boost to GDP boils down to any extra output created by the additional activity that happens in that space. To work out the direct effect on GDP from the cost savings channel, you need to first identify how much of it is cost transfer vs production function change, and then second, estimate the extra GDP produced by the new space ‘freed up’.
Will more WFH affect property markets?
Yes – but how exactly depends on supply elasticities. Let’s begin with the easy bit: the arguments above imply a change in demand along two dimensions. First, a relative decline in demand for city centre housing and a rise in demand for housing farther away. Typically, when living in a city centre, you are getting less space in exchange for a shorter commute. With more WFH the value of a shorter commute declines, and the hassle of living in a small place rises. Second, with people wanting more space at home to work, and lower demand for office space, that implies a demand shift from commercial to residential property.
The harder part is how this translates into price vs quantity adjustments. If supply is perfectly elastic, the supply curve is horizontal and so everything happens via quantities, and prices don’t change. If supply is completely inelastic, the curve is vertical, and any changes in demand show up entirely in prices.
In the very short run, it’s reasonable to assume that supply of most property types is completely inelastic and so prices adjust. Indeed, several papers have shown evidence of clear differences in price moves between city centres, suburbs and towns in the rental market.
But in the longer run, which is more economically relevant here, supply elasticity is a more complex issue. And it’s likely to vary between national jurisdictions and local areas depending on geography, density, planning regulations and local preferences.
On the spatial side, long-run supply in big city centres is likely to be fairly inelastic in response to a negative shock because you can’t ‘unbuild’ houses and turn them into undeveloped land. And there’s a broader issue about housing type: in many places city centre flats are small, not designed for WFH and reallocating housing space into larger units is difficult.
But on the edges of cities, elasticity might be more elastic. For the US, there’s evidence that in some cities supply is fairly elastic on the fringes. But in places where city limits are tightly constrained by regulations, it might be harder for supply to respond. To the extent that supply curves are kinked (inelastic downwards, but elastic upwards), a shift in preferences from city centres to elsewhere would (all else equal) push down on property prices in aggregate because of its asymmetric effect on prices.
On the land-use side, the elasticity depends on how easily existing buildings can be repurposed from commercial to residential, and/or the costs of outright replacement. These costs are partly related to the legal hurdles surrounding land/building use, and partly to refurbishment costs – both the costs of conversion, and broader issues around the attractiveness of converted office buildings as flats. Conversion is economically unviable if the costs of converting exceed the gain from differential values of converting. The bigger the costs, the bigger the price response.
Final thoughts
If it happens, greater post-pandemic WFH could have substantial economic implications. This post is just an attempt to sketch out the economics of three of the simplest macro ones.
But even taking the mechanics above as given, quantifying those will probably requires new tools and/or more work on sparsely researched areas – such as the ‘office space elasticity of GDP’. And there are many other implications not analysed here – labour markets, distribution of income, transport economics, spatial economics and the broader effects on productivity – to name just five.
John Lewis work in the Bank’s Research Hub.
If you want to get in touch, please email us at bankunderground@bankofengland.co.uk or leave a comment below.
Comments will only appear once approved by a moderator, and are only published where a full name is supplied. Bank Underground is a blog for Bank of England staff to share views that challenge – or support – prevailing policy orthodoxies. The views expressed here are those of the authors, and are not necessarily those of the Bank of England, or its policy committees.
