The stamp duty — a tax to be paid on the transfer of certain legal documents — has a long history in England. It was first introduced in 1694 to help finance the war against France. Although initially thought as a temporary solution, the tax was so successful in raising revenue that it eventually evolved into a permanent form of UK taxation. While in recent decades global competition has reduced the scope of stamp duty on the trade of financial assets, stamp duty on transactions of land and property is today more significant than ever. Revenue from the stamp duty land tax (SDLT) has almost tripled over the last eight years, from £2.9bn in 2008-09 to £8.6bn in 2016-17.
The SDLT (commonly labelled ‘stamp duty’) has long been criticised by economists as being inefficient. The central case against it is that it hampers household mobility. The 2011 Mirrlees Review neatly summarises the key argument:
“By discouraging mutually beneficial transactions, stamp duty ensures that properties are not held by the people who value them most. It creates a disincentive for people to move house, thereby leading to potential inflexibilities in the labour market and encouraging people to live […] in properties of a size and in a location that they may well not otherwise have chosen.”
There are some prior studies documenting a negative effect of the stamp duty on housing transactions and mobility, but to date we know little about the nature of affected moves. Specifically, there is no evidence of the relative importance of the SDLT for the functioning of housing and labour markets. In a new article, published in the Journal of Urban Economics (Hilber and Lyytikäinen, 2017), we provide some answers by exploring the impact of the SDLT on different types of mobility.
Before 2014, the UK stamp duty had the odd feature that the tax rate for the whole purchase price jumped at certain thresholds. For example at £250,000 the rate jumped from 1% to 3% of the value (a £5,000 increase in tax liability). We exploit this discontinuity to isolate the impact of the stamp duty from other determinants of mobility. Essentially, we compare mobility rates of otherwise similar homeowners with self-assessed house values below and above the threshold.
The “slab” structure of the SDLT was replaced with a more progressive “slice” schedule in 2014. Our data is from a period before the reform and we utilise the discontinuities of the old system to identify the impact of the SDLT. The results are, however, informative of the effects of SDLT on mobility in the current system too. The reform removed the discontinuities in the tax liability and the bunching of transactions just below the old thresholds. However, it did not alter substantially the overall tax burdens, somewhat lowering them at the bottom end and increasing them at the top end of the price distribution. Thus the tax-induced disincentive to move remained.
Figure 1 below illustrates our research design and key results. We use the British Household Panel Survey (BHPS) to analyse how the 2%-point increase in SDLT affects mobility of homeowners. The BHPS includes homeowners’ own estimates of the value of their home. When we plot mobility rates in house value bins around £250k in Panel A of Figure 1, we find a clearly visible reduction in the rate of mobility when the self-assessed house value exceeds £250k and the tax rate rises from 1% to 3%.
Figure 1: What happens to mobility when stamp duty increases from 1% to 3%?
The BHPS also contains information on the distance of moves and on the main reasons for moving. Interestingly, when we split moves by distance of move (shorter than versus longer than 10 kilometres — Panels B and C in Figure 1), we find a large effect on short moves but no effect on long moves. The stamp duty appears to prevent moves to a more suitable dwelling locally but has little impact on long-distance relocation.
A similar picture emerges when we divide moves into three groups based on survey responses regarding the main reasons for moving (Panels D to F in Figure 1). The stamp duty affects housing- and area-related moves but has little effect on job-related or major life event-related mobility. We should note that job-motivated mobility is overall very rare among owner-occupiers. Nevertheless, all in all our findings strongly suggest that the stamp duty jams the housing market by preventing households from moving to more suitable homes but its effect on relocation of the labour force appears to be quite limited.
A potential explanation for the differential effects is that when moving to a more suitable house locally — say one bedroom more or less — the benefits of moving often exceed the costs only by a narrow margin. Therefore even a small increase in the tax wedge can prevent a large fraction of moves. Longer, often job-related moves — say a great job opportunity in another city — implies large benefits of moving and in turn that the SDLT burden may not be pivotal in most cases.
Our analysis suggests that abolishing the SDLT (or replacing it with an annual tax on the value of land or property) may do little to improve the spatial mismatch of job opportunities and workers. However, it could greatly reduce the allocative mismatch in the housing market via increasing the likelihood of a match between elderly households willing to downsize and young families seeking to expand their housing consumption. We do not claim that abolishing the SDLT could solve the housing affordability crisis. The latter is caused largely by a flawed planning system (see for example Cheshire, 2014, for the central argument and Hilber and Vermeulen, 2016, for estimates of the causal effects of regulatory constraints on house prices). However, it could help alleviate it by putting existing housing stock into more efficient use.
The potential caveat of our research design is that people with higher underlying propensity to move could select into the lower tax group. This could bias our results. It is comforting that we do not find discontinuities in household characteristics linked to the propensity to move. Also, our results are robust to dropping data very close to the £250k threshold where sorting would likely occur. These checks suggest that our research design is valid.
The main scholarly contribution of our study is the analysis of the differential impacts of the SDLT on different types of mobility. Yet, from a policy point of view, the sheer magnitude of the impact of the SDLT on mobility is of course also of great importance. Our estimates come with uncertainty. However, taken at face value they imply that the overall effect is very substantial: our central estimate suggests that a two percentage point increase in the stamp duty from 1% to 3% reduces household mobility by almost 40%. Considering the entire distribution of self-assessed house values in our sample in 2007, our central estimate implies that abolishing (or replacing) the stamp duty could have increased mobility of home owners by around 27% (from 5.1 to 6.5 percentage points). Since 2007 house prices have risen substantially in the UK and the system was reformed in 2014. While the former increased the stamp duty tax burden, the latter reduced it for the majority of homeowners. Overall, the adverse impact of the SLDT on mobility may be of a similar magnitude.
Our empirical findings suggest that the stamp duty induces substantial misallocation of dwellings. This imposes a hefty welfare loss on the society as a whole. We can quantify the welfare loss relative to the additional tax revenue generated by the stamp duty with the help of some simple calculations. These are based on the observation that, for a transaction to take place, the valuation of the buyer has to exceed the valuation of the seller at least by the amount of the tax liability. Therefore, we can assume that each transaction prevented by the tax rate hike from 1% to 3% destroys welfare of the trading partners by somewhere between 1% and 3% of the price of the dwelling. At the same time, each prevented transaction depresses tax revenue. Our calculations suggest that the welfare loss associated with the tax rate hike from 1% to 3% is massive, possibly above 80% of the revenue increase.
The same amount of revenue could be collected with little or no welfare losses through an annual tax on the value of property, or even better, an annual tax on the value of land. In contrast to the SDLT, these taxes do not discourage mutually beneficial transactions. In fact, if designed as local taxes, they would have the additional benefit of providing genuine fiscal incentives to local authorities (and their residents) to release more land for residential development. This would help gradually to solve the housing affordability crisis that cripples the prospects of the young generation, particularly in the most productive parts of the country.
This article was originally published in CentrePiece and is reproduced here with permission.
References
Cheshire, P (2014) Turning houses into gold: the failure of British planning. CentrePiece, Spring 2014, p14-18.
Hilber, CAL and Lyytikäinen, T (2017) Transfer taxes and household mobility: distortion on the housing or labor market? Journal of Urban Economics, 101, p57-73.
Hilber, CAL and Vermeulen, W (2016) The impact of supply constraints on house prices in England. Economic Journal, 126(591), p358-405.
Mirrlees, J, Adam, S, Besley, T, Blundell, R, Bond, S, Chote, R, Gammie, M, Johnson, P, Myles, G & Poterba, J (2011) Tax by Design: The Mirrlees Review. Oxford University Press.
