Assessing the Distributional Impacts of a Land Value Tax Coupled with Income Tax Reform
April 2024 | By Liam Wilkinson. This research note accompanies the modelling found here: Tax Incidence of LVT and Income Tax Reform
Abstract
This research explores the distributional impacts of combining a national land value tax with income tax reforms in Canada. Using income and property ownership data, we find that LVT is largely a regressive tax, particularly when paired with a reduction to income taxes via the enlargement of the basic personal amount (0%) bracket to $88,100. These impacts are substantially mitigated by a flat per household refundable tax credit of $12,700 which leaves 80% of households better off. These findings strongly suggest that a negative income tax or guaranteed basic model may offer a more progressive approach, and ensure that a significant majority of the population enjoy net tax relief while maintaining progressivity.
Introduction
The endeavour to create an equitable and efficient taxation system remains a perpetual pursuit in Canada, particularly in light of the distortions the present system has caused in the housing market[1]. As we strive for fairness, economic dynamism, and a healthy functioning housing market, increasing land value taxes to offset reducing income taxes has emerged as a promising policy combination. However, reforming Canadian landowners' relationship to property taxation remains an incredibly challenging political proposition. This research note aims to uncover the incidence and distributional considerations of combining a land value tax (LVT) with two distinct income tax reforms.
The notion of a land value tax, dating back to the classical economists and Henry George's writings, has resurged in contemporary discourse as a potential remedy to various socioeconomic issues and a critical affordable housing crisis in Canada. Its unique premise, taxing the unimproved value of land irrespective of property improvements, stands in contrast to conventional property taxes. Indeed, many jurisdictions and policy makers forward pairing LVT with tax cuts to the structure component of property taxes[2], otherwise known as building exemptions. This variant is not explored here; rather we explore pairing LVT with revenue neutral income tax reforms. Such a combination could address numerous adverse effects of the current system such as the work disincentive caused by income taxes. This would have the effect of shifting the majority tax burden off individual productive effort and onto communally created land wealth, alleviating income inequality and poverty, and optimising revenue streams for public coffers.
This research builds on the previous publications of Common Wealth Canada[3], and seeks to uncover the distributional implications of such a tax policy amalgamation. By exploring the impacts on different income brackets, socio-demographic groups, and regional disparities, this analysis seeks to clarify the potential shifts in economic burdens and benefits. Additionally, it aims to assess the broader macroeconomic consequences by inferring the likely consequences of the reforms’ incidence.
Understanding the intricate dynamics and trade-offs involved in these proposed tax adjustments is pivotal for policymakers and stakeholders alike. By shedding light on the distributional impacts of coupling a land value tax with targeted income tax reforms, we hope to contribute valuable insights to the ongoing discourse surrounding equitable and effective tax policy frameworks.
The Scale of Potential Tax Reform - Valuing Canada’s Land
In, Natural Common Wealth and Economic Rent in Canada (July, 2023), we produced a national estimate of the total potential net new revenue that could be generated by a land value tax that collects 75% of annual land rents. For detailed review of those estimations, the paper can be accessed in its entirety here.
Using Statistics Canada's National Accounts data[4] on the total land values across the country ($5.824 trillion as of 2022) we calculated the total rents produced by that land, the total tax revenue arising from collecting 75% of that annual rent, the net new revenue after subtracting existing land taxes built into property taxes, and the resulting impact on land prices.
To estimate total available taxable land rents for 2022, we employed the relationship between land prices and land income, known as the capitalization rate. This rate historically fluctuates between 3% and 8% in Canada, with a long-term average of 5.5%. Hence, the paper estimated the total available taxable land rents for 2022 as $320 billion (=$5.824 trillion * 5.5%). However, we noted the vast discrepancies between this figure and other estimation methodologies. For instance, alternate methodologies suggested potential land rents of $593 billion[5], nearly 185% higher than our initial estimate. The paper discussed the potential reasons for such discrepancies, including unmeasured land, imputed rents on government property, or systematic under-assessment of land values.
After considering different proposals suggesting varying percentages of land rent collection, we selected 75% as the land rents collection rate, estimating that such a rate would have generated $242 billion in 2022. This calculation suggested an additional net revenue of $194 billion from the proposed land value tax after accounting for existing property taxes attributable to land value. Finally, we also demonstrated the theoretical relationship between land prices and the taxation rate, finding that a 17% Land Asset Value Tax would capture 75% of land rental value, equating to approximately 4.2% of pre-reform land value or 2.4% of the average residential property value.
Examining the tax incidence
Methodology
While it is fairly straightforward to derive the total tax revenue from the macro level land value data, it is enormously challenging to calculate its incidence across the country. The foremost challenge here is the lack of data. To calculate any household’s LVT burden for any given rate, we must determine the value of any land they own. The accurate assessment of land values is central to the efficacy and efficiency of land value capture but is currently not independently assessed in any jurisdiction outside British Columbia. Therefore we must estimate land values using total property values and the average ratio of land to structure value for any given area. To find this ratio for Canada we can again turn to the National Accounts data, to find the national average; taking the quotient of the value of all land with dwellings and the total property value of dwellings. This gives us an average land to property value ratio of 58%.
While we use this figure in our following estimates it should be noted that it hides enormous variability and should not be taken to calculate the specific incidence on any one household. For instance, the value of land might only comprise 30% of the total property value for a new build in Whitehorse, Yukon; while for a single family home in downtown Vancouver, British Columbia it might comprise as much as % of the total property value.
There are many sources of detailed property value data but to determine the tax incidence we must also know the incomes of the owners of these properties. This is the central limitation in discovering the distributional impacts of an LVT as there are only a handful of data sets that include both income and property value data, and each have serious limitations. Our selected source in this instance is the Survey of Financial Security (SFS), 2019.
The custom tabulation that is the source of all our calculations provided average property values, incomes, and mortgage debts by household national income decile. Given the variability likely hidden behind the national averages (for instance a household in first national decile has an average income of approximately $12,000 while holding over $150,000 in real estate assets) we further break this down by homeowners (retired and not), non-homeowners (retired and not), as well as by region.
*The following spreadsheet references, refer to Sheet “BPA 17% LVT Model” of Tax Incidence of LVT and Income Tax Reform
To calculate the incidence for any given decile, we first calculate the Average principal residence (Column G) and Average secondary (Column K) real estate assets. We then calculate the pre-effect (i.e. pre-LVT introduction) land price by using the product of the property values and the national Land/Property Value Ratio (Column L). Column P then calculated the post-effect land price using the following relationship:
Value of land = land rental value / (land value taxation rate +land capitalization rate)
From there we find the total primary (Column Q) and secondary (Column R) LVT payments.
Given the scope of property value loss in people’s homes (Column X) we also explore an imagined mortgage reform which would eliminate some mortgage debt in proportion to the total loss of property value; in other words, a homeowner whose home value decreases by half would also see their mortgage balance reduced by half. In columns M and T through V we calculate the annual mortgage costs (based on a 15 year mortgage at 4.5%) both before and after this hypothetical mortgage relief program.
LVT
In order to explore how best to structure a revenue neutral tax shift onto LVT, it is important to understand the incidence of LVT on its own, even though it would be accompanied by additional reforms such as those we later explore.
On its face, one might think that LVT would have a relatively progressive incidence given that it is an asset tax and wealthier households hold more assets, however the nature of home ownership violates this assumption. Households in the lower deciles hold higher assets values in proportion to their income therefore the incidence of an LVT reflects this and is highly regressive; falling in much greater proportion on the lowest income deciles (Column Z).
LVT + Income Tax Reform
The merits of LVT stem mostly from the collection of unearned land rents and the economic and social efficiencies that follow. While there are many pro-social and economically efficient ways to use that revenue, the overall political salience of the policy will likely rest on the form and distributional implications of any accompanying tax relief.
In order to explore the fairness of this policy, how the tax burden on families will change, and how the unintended consequences of LVT’s introduction could be mitigated with additional policy instruments, we model two income tax reforms: an increase of the 0% Basic Personal Amount tax bracket to $88,100, and a flat per household refundable tax credit of $12,700.
Increasing the Basic Personal Amount to $88,100
Our first income tax reform imagines that the federal and provincial governments come together to raise the 0% tax bracket (the Basic Personal Amount) equally across all jurisdictions, such that no income taxes are paid on the first $88,100 of income that an individual earns.
This income bracket calculation was modelled by Vivic Research using the SPSD/M to maximise the 0% tax bracket using the LVT revenue calculation of $194 billion as the offset. Full details of this calculation can be found here: SPSDM Simulation Outputs - July 6 2023.
To calculate the income tax reduction (Column S), we simply take the minimum between a household’s current tax burden and the average combined provincial and federal income taxes paid on $88,100. This figure is then added alongside any LVT payments to recalculate the absolute and relative income change for the household.
This calculation has two substantial sources of error. First, we do not know the number of income-earning individuals in any household, therefore we cannot accurately assess the amount of tax reduction the household will experience. We therefore assume for all households, a single-earner which likely significantly underestimates the tax relief for high-income earning households with multiple earners and overestimates the tax relief for low income households with where each earner's income falls within the existing Basic Personal Amounts. Second, income tax rates are different in every province, therefore the amount of tax relief offered by this reform would vary with each jurisdiction.
Despite these errors, a picture emerges nevertheless of the combined distributional impact of these two reforms (Columns AA-AB).
Once again, we see the reforms are regressive (this time to the 8th decile).
As we see above the five lowest deciles simply do not earn sufficient income (and pay the corresponding income tax) for the expansion of the 0% tax bracket to exceed their predicted LVT payment. Only once incomes rise to substantial levels in the 5th-8th decile are earnings sufficient enough for the household to take full advantage of the tax relief. Once however the maximum reduction is reached in the 8-10 deciles, eventually the additional LVT burden on the higher-value homes, once again exceeds the tax relief.
It should be noted however that this average view obscures the picture for low income Canadians significantly. Most low-income Canadians do not have any real estate holdings and thus would not pay any LVT. If we examine the 40% of Canadians who rent (rows 42-51) they are universally left with more income - though once again those with higher incomes are able to take greater advantage of the reform.
Of the two axes for which the data is broken down, homeowner/renter and retired/non retired, we would expect that the incidence of the LVT to fall most on homeowners, while the benefits of the BPA expansion would accrue mostly to the non retired population with its higher average income.
Examining the non retired population that rents (29% of all households) we see that those who rent are on average much better off save for those in the lowest deciles; who do not have sufficient income to take advantage of the expanded BPA yet still, on average, hold some real estate assets subject to the LVT.
For non retired homeowners (46% of all households) the picture looks much the same as the national deciles, given that they hold greater real estate assets and also have higher average incomes.
Meanwhile the retired population (not pictured) holds an above average amount of real estate and a below average household income, therefore the combined effect of this reform falls hardest on them. While the vast majority of retired households who rent (29% of all retired households) are better off, those who own are nearly universally worse off, seeing as much as a 43% decline in income. The reality that such a reform would fall hardest on those who are land asset rich and income poor (and that these are disproportionately retirees) is an obvious challenge. It should be noted however that additional policies such as tax deferment until death or sale, or target tax credits could largely mitigate these impacts though such policies would reduce the amount of revenue available to finance the BPA expansion.
Providing all households with a Refundable Tax Credit of $12,700
*The following spreadsheet references refer to Sheet “RTC 17% LVT Model” of Tax Incidence of LVT and Income Tax Reform
An alternative income tax relief delivery mechanism that we explored was a dividend or refundable tax credit of $12,700, delivered to each household. This would function similarly to the Canada Carbon Rebate (for those provinces who pay into the federal carbon tax) however it would be a flat amount for each household. The value of this credit is simply calculated by dividing all net revenue by the number of households.
This model would also have significant limitations in practice, however it would better allocate the income tax relief in proportion to the unit (a household) that would bear the cost of the LVT implementation. Importantly, because the relief is a fixed amount, it would have the greatest proportional impacts on the lowest income earning household.
This model would leave nearly all households better off, only marginally negatively impacting the wealthiest Canadians. For the lowest-income brackets, this reform would act as a de facto Basic Income, guaranteeing nearly $13,000 to every Canadian household.
Limitations
The challenges in accurately determining the incidence of LVT across different income brackets and regions are significant. The limitations stemming from data availability and the variability in land value to property values highlight the complexities in understanding the tax burden at a household level.
While these results point to the general distributional trends of these reforms, it should be noted that any real world implementation of these policies would, by necessity, likely differ from these simplistic assumptions immensely. One of the key principles of land value taxation is that it will encourage better land use and thus change the current distribution of land ownership, encouraging those with excess land holdings to sell thus increasing the market supply of land. Implementing the reform over many years would enable time for the market and households to adjust while the benefits of such a reform, such as more housing supply and greater economic growth, are realised.
In addition to the practical limitations of the modelling noted above, there are numerous other challenges and distortions that do not necessarily resolve in the macro level data we have presented. In addition there are many reasons to suggest that the full revenue of the reform would not be available to be used for an accompanying income tax reduction.
For instance it is likely a requirement of such a reform that it addresses, in part, some instances of equity loss. For instance, a new homebuyer, having taken on a large mortgage, might see their collateralized asset (the property) diminish in value beyond that of the loan, putting them underwater. This could be addressed with various other tax policy instruments (such as a recent-homebuyers tax credit), but they would likely eat into a significant portion of the available revenue, reducing the available benefit to households - assuming that the reform remains revenue neutral.
This modelling also fails to capture the impact on businesses, many of which hold real estate assets that would be subject to an LVT but do benefit directly from the reforms explored here. For land value intensive businesses this impact could be catastrophic without an accompanying tax reduction.
Conclusion
Combining a Land Value Tax (LVT) with revenue-neutral income tax reforms is a policy option that holds great promise in correcting the perverse incentives of our current system and the excesses it produces. While there are many other tax shifts worthy of careful study, such as reductions to property structure taxes and corporate taxes; this research endeavour aimed to unravel the potential distributional impacts of two such tax policy adjustments, shedding light on the complexities and trade-offs inherent in their implementation.
The modelled scenarios of income tax reforms, including the expansion of the 0% tax bracket and a refundable tax credit, revealed major effects across income deciles, revealing both regressive tendencies and potential redistributive impacts. These findings point to a negative income tax or guaranteed basic model as potentially superior in ensuring the combined reform is largely progressive and does not negatively impact the most vulnerable.
Given that any likely LVT implement and accompanying reform would be constitutionally constrained to being enacted at the provincial level, our next research effort will model these changes provincially. Doing so has several advantages, namely more accurate and available land value data (this is particularly the case in British Columbia where land values are independently assessed).
Additional research should aim to further resolve the incidence of the LVT on households, accounting for regional disparities in the land-to-property value ratio as well as the incidence on agricultural and commercial lands.
In conclusion, the pursuit of equitable and efficient tax policies necessitates a multifaceted approach, considering not only revenue generation but also the broader socio-economic implications. Our exploration, albeit limited by data constraints and inherent uncertainties, provides valuable insights into the potential distributional impacts of integrating LVT with income tax reforms. Further research and policy considerations are warranted to navigate the complexities and ensure the realisation of fair and effective taxation systems that benefit society at large.
Feedback
Feedback, questions and suggestions regarding this work are most welcome and can be sent to info@commonwealth.ca
