In recent years, there has been a great deal of interest in cutting taxes at all levels of government. While tax cuts can serve many different purposes-from increasing economic competitiveness to making the tax system more progressive-much of the recent interest has been stimulated by the claim that tax cuts pay for themselves over time by boosting economic activity and thus government revenues.
At the local level, the surge in tourism and hotel occupancy tax revenues that accompanied New York City's December 1994 cut in its hotel occupancy tax has been used to make the case for cuts in other taxes. However, it is not appropriate to generalize from one tax cut to the next; even cuts in the same tax will have very different revenue impacts under different circumstances.
But it is instructive to take a close look at the City's reduction in hotel occupancy taxes to determine the extent to which recent increases in hotel occupancy and other tax receipts are attributable to economic activity specifically generated by the tax cut. The purpose of this fiscal brief is not to determine whether the City should have reduced the rate on its hotel occupancy tax in 1994-tax cuts do not necessarily have to pay for themselves to be desirable. The much narrower goal is simply to determine how much of the direct revenue loss attributable to cutting the City's hotel occupancy tax was offset by revenue gains from increased economic activity.
A look at the hotel occupancy and room rate data reveals that demand for hotel rooms was increasing strongly and thereby boosting hotel tax receipts well before the City cut its tax on hotel occupancy. Clearly other factors beyond changes in the tax rate influence tourism and City revenues from the hotel occupancy tax. Using sophisticated econometric techniques to distinguish among the different factors underlying the increase in hotel occupancy and room rates, IBO finds that while the City's 1994 reduction in its hotel occupancy tax rate did generate significant increases in economic activity and tax revenues, the tax cut did not pay for itself:
- Increased receipts from New York City's hotel occupancy tax and general sales tax on hotel rooms offset as much as 20 percent of the tax revenue lost due to the cut in the City's hotel occupancy tax rate.
- Adding in secondary fiscal benefits from all City tax sources attributable to the tax reduction's stimulus to the local economy, in the long run the City can expect to regain slightly more than half of forgone hotel occupancy tax revenues.
This study begins with an overview of recent changes in hotel occupancy tax rates and revenues. In the next section, an econometric model is used to estimate the direct fiscal impact of the 1994 City tax cuts, highlighting the extent to which the rate reduction boosted tax revenues through increased hotel occupancy and room rates. The secondary fiscal effects of additional economic activity stimulated by increased hotel stays are examined in the following section. The study concludes with a summary of IBO's major findings and discusses why the City's experience in cutting its tax on hotel occupancy cannot be generalized to cuts in other City taxes. (An appendix detailing the econometric model and its results is available on request.)
An Overview of the Data
The pattern of recent changes in hotel occupancy tax rates, City hotel revenues, hotel occupancy, and nightly room rates suggests that demand for hotel rooms-and tourism in general-began to rise well before the hotel tax reductions.
Hotel Occupancy Tax Rates
Hotel rooms located in New York City are subject to hotel occupancy taxes, which are levied in addition to the State and City's general sales taxes of 8.25 percent. Before July 1986, New York City taxed hotel rooms renting for $40 or more at a flat rate of $2 per night, with lower amounts for less expensive rooms. 1 In July 1986, the City added a variable component to its hotel occupancy tax; although the rate was initially 5 percent of room charges, it was increased to 6 percent in September 1990. (See Figure 1.) In June 1990, New York State imposed its own hotel occupancy tax of 5 percent on hotel rooms renting for $100 or more per night-a statewide tax that primarily affected New York City.
By September 1990, a New York City hotel room renting for $100 was subject to taxes totaling 21.25 percent-the highest rate in the country and over twice the average rate for major U.S. cities (de Seve, 1992). Moreover, taxes on hotels had increased precipitously, rising from 15.25 percent in May 1990 to 21.25 percent in September of the same year.
The difficulties associated with high hotel tax rates were compounded by laws requiring that the taxes be itemized on four separate lines of guests' hotel bills. This requirement made visitors and events organizers keenly aware of the many taxes being levied on their hotel stays, and it is likely that this fueled resentment among visitors. The most formal adverse response to the 1990 State and City tax increases came from meeting and convention planners, who soon began an unofficial boycott of New York City as a location for their events. Figure 1.
Hotel Occupancy Tax Rates - (Rooms Renting for $100 per night)
Independent Budget Office
* The City’s hotel occupancy tax includes both a variable component and a flat tax of $2 per night on rooms costing over $40 per night.
Pressure to reduce hotel taxes led New York State to eliminate its 5 percent hotel occupancy tax in September 1994. Three months later, New York City reduced the variable component of its hotel occupancy tax by one percentage point, to 5 percent. Thus by December 1994, a New York City hotel room renting for $100 was subject to taxes totaling 15.25 percent-the rate prevailing in early 1990.
City Hotel Occupancy Tax Revenues
Total City hotel occupancy tax receipts are the sum of revenues from the variable and flat components of the tax. Revenues from the flat tax on room occupancy are the product of the tax rate (for example, $2 per night for rooms over $40) and the number of nights rooms are occupied. Revenues from the variable portion of the tax simply equal the variable tax rate times the nightly room rate times the number of nights rooms are occupied. A decline in the tax rate, all else being equal, would cause total hotel tax revenues from the variable portion to decline.
Despite the one percentage point decline in New York City's hotel occupancy tax rate in December 1994, total City hotel occupancy tax receipts have actually increased. Figure 2 (see next page) provides some initial insight into what happened. As shown in the figure's top panel, revenues from the City's hotel occupancy tax increased substantially in the two fiscal years following the tax cut, and another solid increase in tax receipts has occurred in fiscal year 1997.
As shown in the figure's lower panel, occupancy rates in New York City hotels have been rising strongly since the early part of fiscal year 1994, both before and after the cut in State and City taxes. In spite of a net decline in the number of New York City hotel rooms during calendar year 1993, the rise in the hotel occupancy rate can be viewed as an indicator of an increase in the number of room nights booked. 2 Although the increase in average room rates before the tax cut was more modest than the increase in hotel occupancy, room rates have also risen sharply in the last two years (City Project, 1997). Together these increases in hotel occupancy and room rates boosted City hotel occupancy tax receipts by more than enough to compensate for the loss of revenues directly attributable to the tax cut.
It is one thing to note that hotel tax receipts rose following cuts in hotel occupancy tax rates and quite another to infer that the tax cuts themselves caused receipts to rise. Both hotel occupancy and room rates in New York City began to increase well before the tax cut, indicating that demand for hotel rooms (and tourism in general) was already rising due to factors other than the tax rate. Moreover, visitors to the City tend to reserve their hotel rooms well in advance, suggesting that the full effect of any change in the tax rate is likely to be evident only after a considerable period of time. Only 17 percent of visitors booked their rooms within three months of their stay, while large conventions often book rooms several years ahead (de Seve, 1992, p. 6).
NYC Comprehensive Annual Financial Report of the Comptroller. FY 1986-1996; PKF Consulting, Trends in the Hotel Industry. various issues.
* Receipts from both the variable and flat components of New York City’s hotel occupancy tax; does not include State hotel occupancy tax receipts or City and State general sales tax receipts.
** City and State hotel occupancy tax rates on rooms renting for $100 per night.
*** Nominal average room rates. Changes in inflation-adjusted room rates follow a similar pattern.
How much of the rise in City tax revenues is attributable to the reduction in the City's hotel occupancy tax rates? A simple examination of the data is not sufficient to address this question. The following section employs more sophisticated empirical analysis to provide some answers.
Estimating the Fiscal Impact
In order to determine the extent to which recent increases in New York City's revenues from its hotel occupancy tax can be attributed to the December 1994 rate reduction, and whether the tax cut on balance has increased or decreased total City revenues from all sources, IBO has estimated the net fiscal impact of the reduction in the City's hotel occupancy tax rate. The first part of the analysis focuses on the rate cut's direct impact on tax revenues derived from hotel occupancy. The calculation of this direct fiscal impact is based on IBO's econometric estimates of how hotel occupancy and room prices are affected by changes in the tax rate. 3 The second part uses a local economic multiplier to examine the indirect fiscal impact of increased economic activity generated by the tax cut's stimulation of hotel occupancy and accompanying visitor spending.
Direct Fiscal Impact
The reduction in the hotel occupancy tax rate has had several direct effects on the amount of tax revenue New York City collects. In and of itself, the rate reduction led to a decrease in hotel tax receipts. Yet because a cut in the tax rate reduces the price of hotel lodging facing buyers, it can be expected to stimulate visitors' demand for rooms, thus increasing hotel occupancy (the quantity of room nights filled by hotels) and perhaps the pre-tax price of hotel rooms per night. Increases in hotel occupancy and room rates in turn cause hotels' taxable receipts to rise. To the extent that taxable receipts increase, City revenues from both the hotel occupancy tax and the general sales tax rise, directly offsetting at least in part the revenue loss from the rate reduction. 4
Effective December 1, 1994, New York City lowered the hotel occupancy tax rate from 6 to 5 percent of visitors' hotel bills (nightly room rate multiplied by number of nights stayed), a one-sixth or 16.7 percent reduction in the tax rate. Neither the flat portion of the hotel occupancy tax, which since September 1980 has been $2 per night for rooms renting for $40 and over, nor the general sales tax rate applicable to hotel rooms was affected. As noted earlier, the City's rate reduction came on the heels of the State's September 1994 elimination of its five percent hotel occupancy tax on rooms renting for $100 or more per night.
In fiscal year 1994, the last complete fiscal year prior to the City and State hotel occupancy tax cuts, the City's hotel occupancy tax receipts (excluding
audits) totaled $126.1 million. An estimated $30 million of this revenue was derived from the flat component of the hotel occupancy tax, with the remaining $96.1 million attributable to the tax's variable component. 5 All else being equal, the one-sixth reduction in the variable portion of the hotel occupancy tax rate led to a direct revenue loss of New York City hotel occupancy tax receipts of $16.0 million per year. 6
This econometric analysis aims to determine the degree to which the City's 1994 tax rate reduction stimulated hotel occupancy and room rates, thereby boosting tax revenues so as to offset the direct revenue loss due to the rate reduction. Revenues from both the flat ($2 per night) and variable (room charges times tax rate) component of the City's hotel occupancy tax would be increased by a rise in the number of room nights filled in New York City hotels, as would the City's receipts of the general sales tax. The total added hotel and general sales tax revenue due to an increase in hotel occupancy caused by the City's 1994 rate reduction may be termed the occupancy effect.
Any increase in nightly room rates caused by the 1994 rate reduction would affect only the variable component of the hotel tax plus, again, the general sales tax. The total of this revenue increase may be termed the room rate effect.
The combination at the end of 1994 of the State's elimination of its hotel occupancy tax and the City's reduction of its tax rate greatly lowered the aggregate tax rate-the sum of City and State hotel occupancy and general sales taxes-on hotel rooms. Thus it is highly plausible that the tax cuts played a significant role in bringing about the recent increases in hotel occupancy and room rates. But, as noted above, these increases began prior to either the State or City tax cut in 1994, indicating the likelihood that other factors also played an important role in stimulating hotel visits and the City's resulting tax collections.
The time-series model. In order to isolate the effects of tax cuts from the effects of other factors, IBO has developed and estimated an econometric, time-series model in which hotel occupancy rates and inflation-adjusted hotel room rates-the model's dependent variables-are each a function of present and past values of the aggregate tax rate and other factors affecting tourism and the hotel industry. In the model, the dependent variables also are affected by past values of each other. Assuming no changes in other variables, a decrease in the aggregate tax rate is expected to cause increases in hotel occupancy and pre-tax room rates.
Similarly, increases in both domestic and overseas economic growth are expected to stimulate tourism and increase hotel occupancy and room rates. As a measure of domestic demand, the model includes a time series of employment growth in the northeastern United States outside of New York City. 7 The Northeast is defined as the twelve states of New England and the Mid-Atlantic regions along with the District of Columbia. The focus on the Northeast stems from a Port Authority survey which found that nearly 75 percent of domestic visitors come from within a 250-miles radius of New York City (Port Authority, 1994, p. 33). 8
To capture the influence of foreign demand, the model includes a weighted index of gross domestic product for eight countries-Canada, France, Germany, Italy, Japan, Spain, Switzerland, and the United Kingdom. Together these countries accounted for over 70 percent of foreign visitors to the New York region in 1992 (Port Authority, 1994, p. 11). 9
The pronounced decline in the incidence of serious crimes committed in New York City in recent years, both in terms of the absolute number of crimes and in relation to the national crime rate, has been linked to the City's attractiveness for tourists. Indeed, surveys of travelers indicate that personal safety and freedom from crime are rated among the most important factors influencing destination choice (Port Authority, 1994, pp. 38-50). A measure of serious crimes committed in New York City is included in the model; decreases in the City's per capita crime rate are expected to improve the public's perception of personal safety, thereby boosting tourism and hotel occupancy.
The cost of doing business is likely to affect the nightly room rates offered by hotels, with increased costs leading to higher real room rates and lower hotel occupancy, all other factors being equal. As a proxy for a more comprehensive measure of the costs facing hotel operators, the model includes a time series of inflation-adjusted payroll expenses per employee of New York City hotels.
Data on the average price of hotel rooms in New York City was adjusted for inflation using the regional consumer price index. The data was adjusted for inflation because real rather than nominal prices influence consumer demand, and because the analysis is interested in the degree to which increases in room prices are a function of underlying structural variables (such as the tax rate) as opposed to general inflation in the local economy. The time series on real room rates was then adjusted for seasonal variation.
Finally, New York City's seasonally adjusted hotel occupancy rate was used as a proxy for the other dependent variable, the quantity of room nights filled. Changes in the hotel occupancy rate are proportional to changes in the quantity of room nights filled for any given number of hotel rooms available, and this analysis assumes that-in the short run-the number of available rooms is not likely to vary appreciably. 10
Monthly data, available for the January 1979 to September 1996 period, was used to estimate the responsiveness of hotel occupancy and real room rates to the various explanatory variables, including the aggregate tax rate measure (the sum of City and State hotel occupancy and general sales taxes levied on hotel rooms). The influence of the explanatory variables is likely to be felt over a long period of time, and up to two years' worth of past values of the variables were included in the estimating equations. Including a large number of past values is also warranted because of how far in advance a large number of hotel rooms are booked, especially by foreign visitors and those attending trade shows and conventions. Because it is likely to take time for public perceptions of crime in New York to change in response to the release of the official crime data (which is itself delayed), the crime rate variable was lagged for an additional two years, so that hotel occupancy in the model is influenced by the crime rate two to four years in the past and not by the contemporaneous value.
The goal of this analysis is to determine the greatest possible impact of the hotel tax cuts on City revenues. Because of this objective, choices in the specification and estimation of the model were made that maximized the estimate of the tax rate coefficient-the measure of the tax rate's influence on hotel occupancy and room rates-in situations where no objective, technical criterion dictated a specific choice. 11
Results of the model. Estimation of the model of hotel occupancy and real room rates generated strong results; about 90 percent of the variation in room rates and 66 percent of the variation in hotel occupancy rates over time were explained by changes in the independent (or explanatory) variables. More importantly, the estimated values of the coefficients-the numbers that measure the influence of the explanatory variables on the dependent variables-generally had the expected (positive or negative) signs and were statistically significant. 12 As expected, the results indicate that contemporaneous and past values of the aggregate tax rate have a negative influence on both hotel occupancy and real room rates, while indicators of domestic and foreign economic growth have a positive influence. Increases in the crime rate measure are shown to reduce hotel occupancy as predicted, and decreases in real per employee payroll costs-the measure of hotel owners' costs of doing business-are shown to increase occupancy.
The results of the estimation also indicate that the relationship between room rates and hotel occupancy differs between the short run and the long run. It is interesting to note that while hotel room rates are negatively correlated with current occupancy rates (higher room rates are associated with lower hotel occupancy), prior values of the occupancy rate in total have a strong positive influence on real room rates. This relationship is not surprising considering how far in advance a large share of hotel rooms are booked. Because hotel operators are able to use computerized reservations systems to track occupancy and alter their offering prices on the basis of changing demand conditions, high occupancy rates combined with information on the number of rooms already booked may signal hotel managers to adjust room rents upward.
For the purposes of calculating the degree to which the direct revenue loss is offset by hotel occupancy and nightly room rate effects, we are interested in the responsiveness of hotel occupancy and room rates to changes in the aggregate tax rate as indicated by the estimated coefficients on that explanatory variable. The results of the estimated model indicate that a 1 percent decline in the aggregate tax rate generates: 1) a 0.34 percent increase in the hotel occupancy rate; and 2) a 0.18 percent increase in real room rates, though the room rate increase is offset in part by the interaction of hotel occupancy and room rates.
Economists would formally term these degrees of hotel occupancy and room rate responsiveness as "inelastic," since a one percent change in the explanatory variable (in this example the aggregate tax rate) leads to a less than one percent change in either of the two dependent variables. Yet the effect on City tax revenue is substantial because the near concurrent State and City hotel occupancy tax cuts in 1994 decreased the aggregate tax rate by over 20 percent.
Less than one-fourth (22.7 percent) of the revenues generated by the combined 1994 hotel occupancy tax cuts, as estimated by the econometric model, can be attributed specifically to the decline in the City's hotel occupancy tax rate. Recall that the City's hotel occupancy tax was reduced by one percentage point, from 6 to 5 percent, while the State eliminated its 5 percent tax on rooms renting for $100 or more per night. While only half of all hotel rooms were subject to the State's 5 percent hotel occupancy tax, these rooms accounted for 68.3 percent of hotel room revenue (de Seve, 1992, p. 16). This information was used to derive the share of revenue offsets attributed to the City tax cut.
Figure 3 summarizes the direct rate reduction, hotel occupancy, and room rate effects on City tax revenues as calculated from the estimated coefficients and the magnitude of the 1994 changes in the aggregate tax rate attributable to the City's rate reduction. The revenues reported in the table are annual figures, stated in terms of fiscal year 1994 hotel occupancy tax revenues. Because the model incorporates up to two years of past values of the aggregate tax rate, the reported figures represent the cumulative effect of rate reduction after two years.
While the rate reduction directly reduces City hotel occupancy tax revenues, a portion of the revenue loss is offset by increases in hotel occupancy and general sales tax revenues due to the occupancy and room rate effects.
The reduction in the City's hotel occupancy tax led to a 1.47 percent increase in hotel occupancy, which in turn generated $1.9 million in City hotel occupancy tax revenue and $1.1 million in City general sales tax revenue. It should be noted that the 1994 rate reduction did not reduce the portion of hotel occupancy tax receipts derived from the flat component of the tax. But the occupancy effect of the rate reduction did increase the number of taxable room nights, thereby boosting hotel occupancy tax revenue. Figure 3.
Direct Revenue Impact of the 1994 Hotel Occupancy Tax Rate Reduction
(In millions of dollars)