- Top of page
In an effort to increase living organ donation, fifteen states passed tax deductions and one a tax credit to help defray potential medical, lodging and wage loss costs between 2004 and 2008. To assess the impact of these policies on living donation rates, we used a differences-in-differences strategy that compares the pre- and postlegislation change in living donations in states that passed legislation against the same change in those states that did not. We found no statistically significant effect of these tax policies on donation rates. Furthermore, we found no evidence of any lagged effects, differential impacts by gender, race or donor relationship, or impacts on deceased donation. Possible hypotheses to explain our findings are: the cash value of the tax deduction may be too low to defray costs faced by donors, lack of public awareness about the existence of these policies, and that states that were proactive enough to pass tax policy laws may have already depleted donor pools with previous interventions.
- Top of page
The shortage of transplantable organs remains a major health care policy issue in the United States. At present, there are over 100 000 individuals on the organ waiting list, a number that has grown over the course of the past decade (1). A variety of different strategies have been developed to address organ shortages, including utilizing organs from marginal donors, improving desensitization techniques for blood group incompatible transplantation, sharing best practice techniques between organ procurement organizations (the Organ Donation Breakthrough Collaborative), large scale education and media campaigns to raise awareness among potential donors, and, for kidney transplants, developing more efficient donor–recipient match algorithms such as paired-kidney exchanges (2–13).
While these strategies have likely been helpful in expanding living and deceased donor pools, growth in demand for organs continues to outstrip growth in supply. In fact, after a steady growth from 1990 to 2004, the number of living donors have become stagnant over the last 5 years. As a result, there has been increased attention on financial incentives to increase donation, particularly those that target living donors (4,14). Outright payment for organs is currently prohibited in the United States: perhaps a reflection of the vibrant debate over the ethics of such “organ markets” (15,16), public support for such policies has consistently been marginal (17). However, there does appear to be support for more limited financial interventions, such as reimbursing living donors for the travel, medical costs, and foregone wages associated with the act of donation (18).
One example of a policy designed with these objectives in mind is Wisconsin‘s 2004 law allowing living donors to deduct up to $10,000 in travel, lodging and lost wage costs accruing from the donation act. While some argued that the policy was unethical in that it had the potential to coerce individuals into donation, tax code breaks for expenses related to living donation have gone on to gain wide support (19,20). Indeed, the policy has diffused to other states and by 2009, 14 other states had passed tax deduction laws with one (Iowa) passing a tax credit law. In addition, another four states are actively considering such legislation (21).
At present, it is unknown whether these tax policies actually do enough to reduce barriers to living donation enough to increase available organ supply. Only one study has attempted to examine the efficacy of these laws on living donation (22). However, this study was limited because outcomes were assessed for only the years immediately before and after the policy was enacted, thereby ruling out the possibility of lagged impacts (i.e. a potential time delay between the legislation's enactment and its eventual implementation and subsequent changes in citizens’ behaviors). Furthermore, the study did not assess differential policy impacts within key subgroups, such as race, gender, and donor–recipient relationship.
This study attempts to quantify the returns to tax policies on organ donation rates at the state level. We improve upon past research by utilizing a large panel data set extending through 2010 that allows us to include more years before and after enactment, thereby yielding more precise estimates of policy effects. We also examined differential impacts by the aforementioned subgroups to test whether the policies have a stronger effect among certain populations.
- Top of page
Table 1 lists the states that passed tax policy laws targeting living organ donation during between 2000 and 2010, as well as the year the policy became effective. Sixteen states passed policies during this period, starting with Wisconsin in 2004. With the exception of Idaho, which offered a $10,000 tax credit, the listed states all allow for up to a $10,000 tax deduction in that year for travel, lodging and lost wages related to living organ donation.
Table 1. States that passed tax deduction/credit legislation and year policy became effective
Table 2 compares descriptive statistics for the 16 states that enacted tax policies against those that did not for years 2000 to 2004, which is the period prior to when the laws were enacted. States with tax policies were similar to those without on all dimensions except income per capita, with states passing tax laws poorer by a small but statistically significant amount.
Table 2. Descriptive statistics
| ||Passed living donation tax policy||p-Value|
|Donors per 100 000 population||2.47 (2.95)||2.64 (1.81)||0.65|
|Per capita income||31 759.10 (5039.66)||29 531.77 (4830.49)||0.00|
|Population (thousands)||6240.31 (7077.95)||5640.97 (4565.44)||0.50|
|ESRD population||8802.98 (10078.31)||8048.19 (6846.26)||0.56|
|% Uninsured||15.49 (4.30)||16.69 (5.84)||0.08|
|% Medicare||14.91 (1.88)||14.59 (1.94)||0.23|
|Transplant hospitals (Number)||7.46 (8.63)||6.80 (4.97)||0.54|
|Conservative Ideology Scale (–3/3 scale)||0.14 (2.06)||−0.19 (1.98)||0.24|
|Number of states||32||15||–|
|Number of observations||160||75||–|
Figure 1 provides a graphical representation of our differences-in-differences regression modeling strategy. The two lines represent the average living donation rates for the states in the control (no law) and treatment (law enacted) groups. Because states in the treatment group enacted laws at different calendar years, the time period was shifted so that the X-axis represents the number of years before and after enactment. Prior to policy enactment, donation rate trends among treatment and control states were similar, though there may have been a slight widening in the gap between the two groups in the 2 years just prior to when the tax laws went into effect. This supports the use of a differences-in-differences strategy for these data, which assumes similar pre-existing trends in the outcome for treatment and control groups. Second, there was no sharp uptick in donation rates the year the policies went into effect (year 0). Although treated states diverge from the control states slightly 2 years after the policy was enacted, this remains smaller than that of the pre-intervention period.
Figure 1. Trends in living donations in states that enacted tax policies versus those that did not. Source: UNOS & OPTN Database. For tax deduction states, the data point at time 0 represents the average donation rate across all states in the year the policy became effective. Time 1 and 2 reflect the postpolicy period, and Time –5 to –1 represent the pre-intervention period. For states that did not pass tax laws, Time 0 reflects the year 2005, the median and modal year tax policies became effective in states with tax legislation.
Download figure to PowerPoint
Figure 2 displays time trends in living donation rates for each state that enacted a tax law. (Control states are not included in this graph.) As in Figure 1, time zero represents the year the policy went into effect, rather than the calendar year. Similar findings emerge. Only a few states show even modest upticks in donation rates the year the policy was enacted, with little evidence of any impacts accruing in later years.
Figure 2. Trends in living donations for each state that enacted tax legislation. Source: UNOS & OPTN Database. The data point at time 0 represents the average donation rate across all states in the year the policy became effective. Times 1–3 reflect the postpolicy period, and Time −3 to −1 represent the pre-intervention period.
Download figure to PowerPoint
Table 3 presents results from our core differences-in-differences regression model. In the model with state and year fixed effects only (column 1), enactment of a tax policy is associated with a 2.91%decrease in the number of donors per capita, although this was not significant at the p < 0.05 level (95% CI ranges from a 13.3% decrease to a 7.5% increase). The inclusion of the time varying state specific characteristics (column 2) and state specific linear time trends (column 3) continues to produce negative and statistically insignificant point estimates for the effect of the tax legislation.
Table 3. Differences-in-differences estimates of the impact of tax legislation on logged living donations per capita
| ||State, Year FE||(1) + State characteristics||(2) + State specific trends|
|Tax legislation (=1)||−0.0291||−0.0372||−0.0502|
| || (−0.133, 0.075) || (−0.130, 0.056) || (−0.185, 0.085) |
|Ln (Per capita income)|| ||0.552||1.251|
| || ||(0.703)||(0.801)|
| || || (−0.864, 1.967) || (−0.362, 2.864) |
|ln (ESRD population)|| ||0.0786||0.293|
| || ||(0.731)||(1.493)|
| || || (−1.393, 1.551) || (−2.712, 3.297) |
|Conservative ideology scale|| ||0.00157||0.0112|
| || ||(0.0118)||(0.0103)|
| || || (−0.0222, 0.0253) || (−0.00957, 0.03203) |
|% Uninsured|| ||−0.00109||0.00228|
| || ||(0.00709)||(0.00680)|
| || || (−0.0154, 0.0132) || (−0.0114, 0.0160) |
|Transplant hospitals (Number)|| ||−0.00605||−0.00188|
| || ||(0.00473)||(0.00538)|
| || || (−0.0156, 0.0035) || (−0.0127, 0.0090) |
|% Medicare|| ||0.00598||0.0127|
| || ||(0.0141)||(0.0164)|
| || || (−0.0224, 0.0343) || (−0.0203, 0.0457) |
|Ln (Population)|| ||−0.524||1.560|
| || ||(1.051)||(2.080)|
| || || (−2.639, 1.592) || (−2.626, 5.747) |
|Number of observations||466||466||466|
One issue with our differences-in-differences specification is that it combines the states that passed tax deductions with the one state (Iowa) that passed tax credits. In principle, these two policies differ in terms of their cash value and could have markedly different impacts on donation. To get at this, we estimated models that did include Iowa, and found markedly similar results. This is not surprising given that Iowa did not experience any bump after passing their tax credit law (see Figure 2).
In terms of the control variables, we do not find any statistically significant coefficients. The signs on state income, ESRD population, political preferences, and% Medicare population are positive and, counter-intuitively, the estimate on the number of transplant hospitals is negative. The signs on the other variables vary depending on the specification. The nonfindings with respect to the control variables were likely driven by our use of state fixed effects—if there is only a small amount of year-to-year variation in these variables within a given state, state dummy variables will soak much of this up and leave little variation with which to estimate precise coefficients on the controls. To address this we estimated specifications without state fixed effects. The substantive results around the tax policy laws were unchanged. However, we found now that percent uninsured and percent on Medicare were both negatively and statistically significantly associated with donation rates. State per capita income, in some specifications, was positively associated with donation rates (these results are available upon request). Given our focus on the impacts of the tax policies, we opted to focus on the specifications with state fixed effects given our desire to better recover causal effects.
Table 4 includes both contemporaneous enactment as well as one and two year lags of the tax policies in order to test whether there is a time delay between enactment and changes in donation rates. The rows display the estimated coefficient for the tax policy main effects variable. We do not find any evidence of lagged effects. The point estimates for the contemporaneous and two year lags continue to be negative and insignificant. The point estimate on the 1 year lag is positive, implying a 3% increase in living donations per capita as a result of tax policies, although it is not statistically significant.
Table 4. Differences-in-differences estimates allowing for lagged policy effects
| ||State, Year FE||(1) + State characteristics||(2) + State-specific trends|
|Tax legislation (=1)||−0.0422||−0.0463||−0.0613|
| || (−0.118, 0.034) || (−0.123, 0.031) || (−0.180, 0.057) |
|Tax legislation, 1 Year Lag (=1)||0.0352||0.0307||0.0244|
| || (−0.0343, 0.1047) || (−0.0391, 0.1005) || (−0.0658, 0.1146) |
|Tax legislation, 2 Year Lag (=1)||−0.0271||−0.0290||−0.0328|
| || (−0.134, 0.079) || (−0.140, 0.082) || (−0.159, 0.093) |
|Number of observations||466||466||466|
Table 5 presents results stratified by relationship to donor and Table 6 presents results stratified by race and gender. We do not find any significant effects for any of these subgroups. We also found no impacts on deceased donation rates (last column, Table 5), suggesting that this was not crowded out by reducing barriers to living donation. Finally, we examined whether the impact of tax policies were higher for states with lower income per capita and did not find any evidence of a significant interaction effect between the tax policy dummy variable and state income per capita. We also assessed whether restricting the control group to states that bordered those that enacted tax legislation would produce changes in our substantive conclusions, and our results remained very similar regardless of how the control states were chosen. (These results are available upon request.)
Table 5. Differences-in-differences estimates stratified by donor relationship
| Panel A—Donor Relationship ||Living, Related||Living, Unrelated||Anonymous||Deceased|
|Tax legislation ( = 1)||−0.024||−0.128||0.155||0.012|
| ||(−0.150, 0.111)||(−0.319, 0.062)||(−0.302, 0.613)||(−0.052, 0.075)|
|Number of observations||466||457||466||421|
Table 6. Differences-in-differences estimates stratified by race and gender
| Panel A: Stratifying by Gender ||Men||Women|
|Tax legislation ( =1)||−0.129||−0.030|
| ||(−0.280, 0.024)||(−0.167, 0.108)|
|Number of observations||386||386|
| Panel B: Stratification by Race ||White||African-American|
|Tax legislation ( =1)||−0.038||−0.127|
| ||(−0.134, 0.058)||(−0.319, 0.065)|
|Number of observations||345||352|
- Top of page
While the establishment of formal markets for organ sales remains politically unpalatable in the United States, 16 states have now implemented tax policies to help incent living organ donation by reducing travel and other incidental costs faced by potential donors. With the exception of Iowa, which offers a tax credit, these states all now offer up to a $10 000 tax deduction for expenses related to the donation act.
In this study, we assessed the impact these tax policies on living donation rates. We used a differences-in-differences strategy to compare the pre- and postlegislation change in donors in states that enacted legislation against the same change in those states that did not, and found no statistically significant association between the passage of tax laws and living organ donation rates. This finding was robust to the inclusion of a variety of control variables, state specific time trends, and allowing for lagged effects. If there were an effect, our estimates suggest it to be small relative to the nationwide gap between organ supply and demand: the highest upper bound of our estimated 95% confidence intervals implies only a 10% increase in living donations accruing from tax policy enactment. We did not find any evidence of any differential impacts within key subgroups either.
There are several potential reasons why these tax laws may not have impacted organ supply in an appreciable way. First, and perhaps most likely, the actual financial return from these laws is quite small. For example, a tax deduction of $10 000 for a family of four at the median income in the state of Wisconsin translates to an actual cash value of just over $600 (22). In contrast, according to one study, the financial burden for a living kidney donor in the US (including transportation, lodging, lost wages, and medical costs) ranges from $907 to $3089, depending on the type of surgical approach (32). Thus, in most cases, the value of the tax deduction represents only a fraction of the true total costs faced by living donors. Along these lines, policies that do more to reduce financial burdens—such as increasing the value of tax deductions or moving from tax deductions to tax credits, which for the same dollar amount would be more valuable and, if refundable, can potentially reduce a payer's tax liability below zero—may have a larger impact.
Second, it is also possible that few people actually knew about the policy, therefore being unable to take advantage of it. Although we do not have survey data to fully support this hypothesis, there is some anecdotal evidence to support this speculation. One author (ASV) contacted different state Donate for Life campaigns to verify the presence of tax incentives, but representatives in only a few states were aware of the existence of these policies. Another author (AV) noticed a lack of knowledge about eligibility for tax incentives within a clinic population being evaluated for living donation, in particularly among the most educated and informed donors. Even with knowledge of the law, potential recipients may face high opportunity costs with respect to the procedures and paperwork to obtain tax deductions, which may be burdensome. This could also blunt the impact of well-intentioned programs.
Third, it may be that states that were proactive enough to pass tax policy laws had already been active in other ways to increase donation rates. This would have depleted their pool of potential donors prior to the enactment of the tax policies, making the law less effective. Along the same lines, tax policies may be more effective when combined with other policies, such as donor awareness campaigns.
Finally, it is possible that any positive policy effects were offset by citizens’ negative reactions to what could be construed as a financial incentive. That is, although some individuals may have been swayed to donate, others were “turned off” by the financial incentive and chose not to donate when they otherwise would have voluntarily done so. The negative point estimates, though statistically insignificant, are suggestive of such a dynamic. Along these lines, it may be that this particular form of altruistic behavior is unrelated to finances: individuals who want to donate will commit to that regardless of cost, especially if they know and care for the recipient. That we find no impacts on anonymous donors either suggests that finances may not matter for any altruistically minded individual, though we cannot exonerate the low cash value of the tax policies as an explanation for this particular result.
There are several limitations to this study, many of which motivate further research. First, it may be that policy effects do accrue, but only several years after the policy is in place. While we explore lagged effects up through 2 years, future research could consider even longer lags. In addition, as alluded to above, tax policies may work interactively with other efforts to increase organ donation, such as provision of paid leave for living donors or public education. We did not include other policies or strategies in our model. Future work that incorporates additional policies would be helpful in designing a multi-pronged package of polices that would have the greatest positive impact.
We do not have data on donation rates by level of education. Educated individuals may have been more likely to have heard about the tax policies and therefore more likely to react. Individual level data with detailed donor demographics and socioeconomic status may be useful in distinguishing who is most likely to respond to financial incentives, which would be useful information to better publicize these tax benefits and improve future policy design. Lastly, and perhaps most importantly, we have only evaluated whether tax policies increased live donations. However, tax deductions/credits may have had other effects among donors, such as improving their well-being. Evaluating additional outcomes such as quality of life and financial hardships would be a promising area for future research.
Despite these limitations, our study makes several important contributions. We utilized a transparent natural experiment study design and a variety of specification checks to recover some of the first estimates of the impact of financial incentives on organ donation rates in the United States. We then extended the scant existing literature on this topic by examining whether policy impacts accrue in the future (if not in the present) or only within key sociodemographic subgroups. Finally, we provided detailed hypotheses as to why we find no policy impacts, each of which inform both future research and policy design.