In collaboration with Dane Lindholm and Malachi Rhoden
One policy resource that governments have is the power of taxation. Selectively making certain areas tax havens creates incentives for development to occur in these designated areas. They can stimulate spending and development; an important concept when analyzing how to meet certain urban-planning goals.
During a prior stage of inequality in the late 1800s, social reformer Henry George proposed a new solution to motivate economic development. He proposed to “abolish all taxation save that upon land values” (George 1879). Since the financial crisis of 2008, much of the world has had difficulty incentivizing landowners to invest and construct on undeveloped land. Government agencies, the same that influence tax rates, must now consider strategies in regards to motivating owners to build upon their land. George’s proposal to cut taxes on developed land would motivate landowners to build on their properties as a source of increasing cash flows through said tax break, and would therefore penalize landowners who choose not to build on their land. A reduced or nullified tax on developed land would increase productivity and could spur economic growth within a municipality.
Although models for the relationship of property taxes and the supply of housing have been extensively theorized since Henry George’s proposal, sufficient data of empirical evidence has been difficult to come by. Fortunately, in May of 2007 Teemu Lyyikäinen published an article using mathematical models to analyze the empirical results of a Finnish tax reform from 2001 until 2007. The reform allowed municipalities to create a variation in the tax rates for undeveloped residential land, and contained quality data to estimate the impact of property tax incentives on housing construction. The goal of this reform was to give Finnish municipalities a method of encouraging housing construction by implementing tax incentives once undeveloped land was zoned for housing. The purpose was to investigate how implementation of a three-rate tax system that taxed undeveloped land at a higher rate than developed land, specifically zoned land, commercial buildings, or permanent dwellings, impacted housing starts and housing density within the region’s municipalities.
Lyyikäinen’s (2009) article introduces the Finnish property tax system and continues with an explanation of the theoretical mathematical model and the empirical results of this model. Lyyikäinen uses the data from the ALTIKA database, which provides data on housing starts, annual population, annual supply of buildings, and housing prices from Q1 of 1998 to Q3 of 2006. The total number of Finnish municipalities that implemented a three-rate system during the 2001 to 2007 period was 118 out of 416. This data was reinforced with property tax rates supplied by the Association of Finnish Local and Regional Authorities (Kuntaliitto). The final property tax rates that the Finnish government implemented in 2007 after the results of the reform analyzed were: (1) a general property tax including land zoned and commercial buildings of 0.50 to 1.00 percent, (2) a property tax on permanent dwellings of 0.22 to 0.50 percent, and (3) a property tax on undeveloped residential land lots of 1.00 to 3.00 percent.
The theoretical model used was an alteration of Turnbull’s (1988) model, as Turnbull’s definition of “landowner value” contained only essential assumptions about urban rent prices from capital for construction. The model was further modified to represent the Finnish property tax system. This dynamic model calculates the present value of the parcel of land by adding the land rent in nonhousing use and the market rent based on the amount of capital used in construction. The model then subtracts off the effective building tax rate based on capital used in construction, the tax on land based on the raw site value, and the effective postdevelopment land tax. This calculation, however, is conditional on structural density and timing of development. Economically, it would be estimated that the theoretical observed effect of the property taxes on start time and density would depend on the incremental returns of a new housing investment.
The empirical model follows a count data analysis similar to the Plassmann and Tideman (2000) model that analyzed the two-rate property tax system. Lyyikäinen (2007) estimates the fixed-effects Poisson (FEP) model proposed by Hausman et al. (1984), who assumed that variance equals mean under weak assumptions is consistent with an estimated Beta parameters. Lyyikäinen specifically chose this model because it allowed for them to not specify a start time for development (no dependence between acquisition and development), in addition to accommodating the volume of buildings started. This is important economically since the theoretical model does not account for the number of buildings started, so the definition of “developed land” could be subjective. In such a case, landowners could potentially build anything to benefit from a lower tax rate. The final model used had two objectives: (1) to estimate the number of housing starts (timing of development) due to the effect of property taxes and (2) to estimate the volume (development density) of housing starts due to the effect of property taxes.
Due to the use of actual data on housing starts to measure construction activity, as well as a large sample size contributing to the estimates of interest, the quality of the Finnish data used in this study is high. When comparing the municipalities that maintained a two-rate tax system against municipalities that adopted a three-rate tax system at some point during 2001 to 2007, the results verify that a three-rate tax system promotes more single-family housing starts. This seemed particularly effective for any municipality that adopted the three-rate tax system in 2003 or 2004, while the evidence is less conclusive for the municipalities that adopted the three-rate tax in 2006 or 2007. One possible explanation of this discrepancy is that the three-rate tax system was implemented later in the year in 2006 and 2007, so the developers had less time to react to the new systems.
To further analyze the data, two FEP regressions were used to calculate the effect of both single-family housing starts and all housing starts on both a three-rate tax dummy variable and the control variables. The control variables used in the study included housing prices, housing stock per capita, province effects per year, and common quarter dummies. The first regression estimated the effect of property taxation with a dummy variable indicating the tax rate system of the municipality, while the second used a complete set of tax rates. The first regression will not suffer measurement errors in tax rates, but contained less detail than the regression with a complete set of tax rates.
The results of the first regression suggest that single-family homes are more susceptible for housing starts with a three-rate tax system than all housing starts. The coefficient of the three-rate tax system for single-family homes (0.121) is highly significant and suggests that the three-rate tax system increases single-family homes starts by 12.1 percent per unit of time. In contrast, the volume per start coefficient is insignificant to indicate that the density of development is not impacted by the three-rate tax system. The coefficient of the three-rate tax system on all housing starts is 8.7 percent, which is less than single-family housing starts, and the volume per building was −2.7 percent, which is again, statistically insignificant. Although the volume per building number is insignificant, the negative sign could hint that the three-rate taxation could lead to slightly lower densities. The reason that single-family housing starts may be more responsive to tax incentives could be that single-family homes are more impacted by land taxes than multiunit housing.
The results of the regression with the complete set of tax rates follows a similar trend to the tax system dummy regression, by demonstrating a significant increase in housing starts due to the new tax rates, but not giving sufficient evidence of an increase in housing density. One benefit of the large variety within the sample size for this regression is that there is significant variation in the differences of tax rates over time within municipalities. Thus, the result of this regression parallels the theoretical model by indicating that higher taxes on undeveloped land incentivize development.
The overall empirical results of the Finnish data suggest that higher taxation on undeveloped land, as compared to developed land, has a positive effect on single-family housing starts, but not on the density or volume of the development. While all-housing starts seem to also be affected by the higher taxation on undeveloped land, the results seem to be less responsive than only single-family starts in the Finnish taxation example. A further conclusion is that development decisions do not impact the base land tax. This signifies that if municipalities have a uniform land tax rate, then development is not affected by land taxes.
The conclusions of this case illustrate how tax rates can stimulate spending within an economy. This relates to Mian and Sufi’s (2014) novel, “House of Debt”, as it confirms consistent evidence that spending or consumption can stimulate an economy that is hurting, or in a recession. For instance, the conclusions from the Finnish three-rate taxation test show that higher land taxes could incentivize landowners to develop on their land to avoid these increased costs. This, in turn, would create more jobs in the construction sector and promote increased consumption within an economy. The increased spending within an economy could bring said economy out of a recession, and shows how governments, or municipalities can play an essential role in revitalizing an economy.
While the results of Lyyikäinen’s (2009) paper help to provide empirical evidence on a theoretical idea that has existed since the late nineteenth century, there are an identified few weaknesses in the paper’s results. First, the results of the Finnish example are from a single period in time (2001 to 2007). Therefore, the reactions to increased development could reflect the market conditions for the time the study occurred, instead of from the taxation reform. This could be the reason why the implementation of the three-rate system in 2003 and 2004 had much stronger results than the implementation of the three-rate system in 2006 and 2007. A second weakness of the paper is that Lyyikäinen used an assumed fixed tax rate difference between predevelopment tax rates and postdevelopment tax rates in the theoretical model. It would be interesting to repeat this study by testing different pre- and post-development tax rates with both larger and smaller differentials. To see if the results of the study remain consistent and could be imposed as a tax incentive globally, one could replicate this research in an economy indicating signs that it is headed toward or already in a recession. If taxing undeveloped land at a higher rate has a correlation to increased spending due to construction and other areas of developmental spending, then it could potentially help to bring economies out of a possible recession.
-
1. Ideally, according to Lyyikäinen (2009), what would be the expected outcome of imposing a three-rate tax system on undeveloped land versus a two-rate tax system?
a. Landowners would cease building on the land in order to decrease the amount since building tax rates are highest
b. Landowners would have an incentive to build on the land since undeveloped residential lots would have the highest tax rate
c. Landowners would cease to build houses since they are heavily taxed
d. Land owners would have an incentive to build commercial buildings since they have the lowest tax rate
Explanation: The correct answer is (b): The lowest tax rate would be on residential properties (permanent dwellings) at a rate of 0.22 to 0.5 percent which would rule out (b) and (d). The middle tax rate would be on zoned land and commercial buildings at a rate of 0.5 to 1 percent, making (a) false. Option (b) is the only option left and is correct since undeveloped residential lots would have the highest tax of 1 to 3 percent%, so landowners should have an incentive to build on that land.
-
2. According to Lyyikäinen (2009), what would be the long-term effect of the adoption of a higher property tax rate on undeveloped land?
a. The effect of adopting a higher tax on undeveloped land should start to decrease over time
b. The effect of adopting a higher tax on undeveloped land should start to increase over time
c. There would be no change on the effect since the availability of land would remain stable
Explanation: The correct answer is (a): Higher taxes on undeveloped land would serve as an incentive to develop on land. This would cause landowners to develop on their land immediately to benefit from the tax incentive. As a result, there would be faster development on land and, overtime, there would be less land to develop on so the overall effect would start to decline. According to the scholarly research, the effects of the tax incentives start to weaken due to approaching toward an equilibrium.
-
3. According to Lyyikäinen (2009), increasing tax on undeveloped land had the greatest positive effect on which of the following type(s) of real estate properties?
a. Zoned land and commercial building
b. Single-family housing units
c. Multi-family housing units
d. All housing units
e. (a) and (b)
Explanation: The correct answer is (b): Based on the data presented by Lyyikäinen (2009), 444 of single housing family houses that were started were due to tax incentives created by the three-rate tax system. This was a 3-percent increase in total single-family housing units started in 2005. When analyzing the tax system dummy regression in the article, the single-family homes had a 12.1-percent increase in housing starts, while all housing starts had only an 8.7-percent increase. Similarly, in the Poisson tax rates regression, the single-family homes had a 5.5-percent increase in development starts compared to 4.6-percent increase in starts for all housing. Therefore, single-family housing starts were concluded to be the greatest observable increase due to the implementation of a three-rate tax system.
References
George, H. (1879), “Progress and Poverty: An Inquiry into the Cause of Industrial Depressions and of Increase of Want with Increase of Wealth; The Remedy,” Modern Library: New York.
Hausman, J., B. H. Hall and Z. Griliches (1984), “Econometric Models for Count Data with an Application to the Patents — R & D Relationship,” Econometrica 52, 909–938.
Lyytikäinen, T. (2009), “Three-rate property taxation and housing construction,” Journal of Urban Economics 65, 305–313.
Mian, A. and A. Sufi (2014), “House of Debt,” University of Chicago Press: Chicago, USA.
Plassmann, F. and T. N. Tideman (2000), “A Markov Chain Monte Carlo Analysis of the Effect of Two-Rate Property Taxes on Construction,” Journal of Urban Economics 47, 216–247.
Turnbull, G. K. (1988), “Property Taxes and the Transition of Land to Urban Use,” Journal of Real Estate Finance and Economics 1, 393–403.