mailto:firstname.lastname@example.orgMore articles by David Lawyer
Are the roads, streets and highways that we drive on in the United States subsidized by the government, or do the taxes on gasoline and license fees pay all (or most all) of the highway costs? And if highways (includes roads and streets) are subsidized, how much is the subsidy? We are talking here about the costs of construction and maintaining highways, including highway patrols, etc. The highway costs are of course only part of the costs of highway transportation which also includes the following: the costs of purchasing and maintaining motor vehicles, costs of motor vehicle fuel (before taxes), costs of parking structures, etc.
If one searches the Internet for the answer to the question of highway subsidy, one finds a number of sites that claim that there is high subsidy for highways. The implication of this is that since highways are heavily subsidized, it's OK to also subsidize other modes of transportation such as mass transit and Amtrak. The fallacy of such reasoning is that one wrong doesn't justify another (or two wrongs don't make a right). If highways are subsidized and this subsidy is wrong, it doesn't justify subsidy to other modes of transportation which may also be wrong.
A couple of websites claim that there is little or no subsidy to highways. The first on is Are Highways Subsidized? by Randal OToole which uses the same method I used to estimate subsidy and obtains the same results. His article goes into the history of automobilisation of America and portrays the auto as bringing great social benefits at modest social cost. I don't think so as the environmental costs were high! But to pursue this would be off topic.
The second website is US Highways & Streets: Revenues & Expenditures by Wendell Cox. I believe that Cox's analysis is biased and understates subsidy. See Discrepancy in Wendell Cox's Table. However, the claims made by others that highways are heavily subsidized are even more erroneous (in the opposite direction).
So is there an honest answer to the question of highway subsidy? This report will show that while there is no exact answer to this question, it looks like there is significant subsidy to large trucks using the highways but it's not clear if there is subsidy to the automobile for their use of the highways. As you will soon see, the situation is far from simple.
Before the automobile in about 1900, the huge highway system of the United States, consisted of about 2 million miles of dirt roads (with a few exceptions) and was about 10 times the length of the railroad system. These roads were almost all 100% subsided by the general taxpayer. Except that major bridges charged tolls to cross rivers.
Starting with the Lancaster (Pennsylvania) Turnpike in 1794, there had been a large number of toll roads, but by 1860 they had mostly failed due in large part to competition from canals and railroads. The turnpike era from 1800 to 1830 saw the construction of many better quality unpaved toll roads known as "turnpikes". From 1845 to almost 1860 was the era of plank toll roads introduced from Russia. Vehicles rolled along on thick wood planks but the exposed wood soon deteriorated. Several thousand miles of toll roads were built. See Transportation Revolution and Transportation (by Bigham)
But even during the toll road era most of the roads were still free (100% subsidized) but in poor condition. The toll roads were 100% pay and usually in good condition. But remember that due to the typical poor condition of the roads and the slow speeds (usually a few miles per hour or less) and low population, there was little travel (from the perspective of today).
The last two decades of the 1800's saw the start of the "good roads movement which was mostly due to support of farmers and bicyclists who wanted the government to pay for good roads. As a result, some states created highway departments with highway financial aid from the states. New Jersey was first in 1881 but it wasn't until 1917 that all state governments were aiding (subsidizing) highways. and Transportation (by Bigham) p.103
At first, the automobile use of highways was 100% subsidized, but that started to change in 1910 when there were almost a half million automobiles in the U.S., and New York state started charging automobile registration fees. Soon all states had such fees and in 1919 (when there were almost 8 million autos) Oregon imposed taxes on gasoline to finance roads. By 1929, with 23 million autos, all states had such fuel taxes. See Locklin, 1947, p.670 Not to be outdone, a federal tax on motor vehicle fuel was introduced in 1932. See Transportation (by Bigham) p.105.
But even after these user charges were in place, minor roads (such as residential streets) were still mostly paid for by local governments (city and county). However in the late 20th century, highway user charges started to be used for non-highway purposes, including the purposes of reducing the national debt and to support mass transit. In a couple of cases these diversions (a reverse subsidy of sorts) exceeded the subsidy from property taxes, etc. for local roads. This provided a rationale for some to claim that highways were no longer subsidized. Even when local streets are subsidized, one may argue that it's not really a subsidy Is Subsidy to Highways Always a Subsidy? and Allocation of Costs to Trucks.
Highways are financed by user revenues (gasoline tax, vehicle registration fees, tolls) and by non-user taxes (property tax, sales tax, and income tax). Some of the money comes from surplus user revenues from years past, including interest earned on them. Today, more and more highway financing is being paid for by going into debt and issuing bonds. The situation is further complicated since about 20% of highway user revenues are being diverted for non-highway purposes such as mass transit. So how would one define user fees paid by highway users? Does it include the money paid for gasoline taxes, but diverted to non-highway uses such as mass transit? Or is this not really a diversion but an excise tax imposed on highway users for special purposes. This is discussed in detail in the next section. Another problem is debt financing. How much of this debt will be paid back by highway users and how much will be paid back by non-users?
One complication is that part of the money obtained from gasoline taxes at the fuel pump are spent for non-highway purposes. In 2005, about 19% of user revenues (including registration fees and tolls) were diverted for nonhighway purposes with over 45% of the amount diverted going to subsidize mass transit. This contributes to a shortfall in revenues needed for highways and it's made up for by using property taxes, general revenues, and other taxes to pay for highways. Should these non-user revenues count as being used for highways or are they indirectly flowing to mass transit and other non-highway uses?
For 1995, the total amount of non-user revenues used for highways just happened to be about equal to the amount of user-revenues diverted for non-highway purposes (such as mass transit). Is there subsidy here? Well, there certainly is subsidy for mass-transit, etc. But are highways being subsidized? Highway user-revenues are certainly sufficient to pay for highway costs if they were used for that purpose, so one may claim that there isn't any subsidy to highways. But part of the money users pay flows to support non-highway purposes.
But for 2005 (unlike 1995) non-user revenues were about double that of the diversion thus implying at about a 13% subsidy to highways. Furthermore, its not exactly clear that this 13% is really a subsidy as is discussed later on in Is Subsidy to Highways Always a Subsidy? and Allocation of Costs to Trucks.
Why the big change from 1995 to 2005? In 1995, the federal government was diverting substantial amounts of gasoline taxes (4.3 cents per gallon) to reduce the national debt (deficit) but this ceased in 1997. See Addendum to the 1997 Federal Highway Cost Allocation Study Final Report.
This diversion in 1995 increased the total diversion to the point where it was about equal to the non-user revenues from property taxes, etc. collected that year. Repealing the diversion, enacted in 1993, resulted in positive net subsidy to highways, the normal case before this diversion started.
For summary statistics on which the above discussion is based, see "Funding for Highways and Disposition of Highway-User Revenues, All Units of Government", Table HF-10 in "Highway Statistics" (annual) by the Federal Highway Administration. See hints for finding Table HF-10 on the Internet.
If one neglects to consider the diversions from gasoline tax, etc., then one can look at table HF-10 for 1995 and claim that about 25% of revenues from taxes came from property taxes, general funds, and other taxes. While this is true, it's not really subsidy since it's equal to the amount diverted from user-revenues. We are asking the question: Do user-revenues (including revenues diverted) fully cover highway costs?
Another way to look at this is to consider the diversions from gasoline taxes as subsidy to mass transit, and subsidy to other non-highway purposes. In other words highway users are subsidizing non-highway projects but at the same time, general taxpayers (property, sales, and income taxes) are subsidizing highways. The question being asked is: what is the net subsidy to highways, which is equal to the subsidy to highways less the subsidy paid by highway users for non-highway purposes.
An example which illustrates why the real subsidy should be only the net subsidy is as follows. Suppose that all taxes on airline tickets were used to pay for highways and conversely: all highway user taxes were used to pay for the operation of the airways and airports. Furthermore, suppose that these costs were equal: the user tax for airline tickets is equal to the taxes paid on gasoline, tolls, etc. Then one might claim that the airways are 100% subsidized since highway users are paying for them. Likewise, one might claim that highways are 100% subsidized since air travelers are paying for them. But actually, the user charges for each mode are sufficient to pay all the costs for that mode and neither the airways nor the highways are being subsidized. Each system pays the other the same sum of money which is equivalent to no exchange of money at all. If I give you $1000 and you give me back $1000, it's like no exchange has taken place (we each have the same amount of money as before). Of course, the actual subsidy situation is much more complex.
While I argue above that it's the net subsidy which counts, there is the opposite point of view described as follows: If one believes that the highway user should be subject to "sin taxes" like those leveled on tobacco and alcohol, then after paying the sin tax, there is still a tax to be paid to finance highway construction and maintenance. From this perspective, one should count the total subsidy to highways and not the net subsidy. Then claims of heavy subsidy to highways would be justified.
The key question here is: Is using the highways a sin? Is using highways hurting the environment? Well, yes, but lots of other activities are also hurting the environment but avoid sin taxes. Let's first look at energy efficiency of the various modes of transportation: Transportation Energy Data Book, Ch. 2. It turns out that the energy-intensity of the automobile is roughly equal to that of mass transit. So if using the highways is damaging the environment, so is using mass transit which is being subsidized by highway users and not paying any "sin taxes". Why aren't the "sin taxes" being levied on transit users as well? Furthermore, since most other activities such as heating and air-conditioning ones home, is consuming fuel and doing environmental damage, why are not sin taxes being levied on all energy consuming activities? Good question. Perhaps what we need are sin taxes on all fuels.
So it seems that levying sin taxes on just motor vehicle use of highways is unduly discriminatory. So thus the subsidy to highways should probably be the net subsidy.
It would make the analysis much simpler and clearer if all of the user revenues for highways were spent only on highways and the funds used for non-highway purposes were obtained from non-highway revenues. But that's not the way it's been set up by government.
NET SUBSIDY TO HIGHWAYS (in billions of dollars) (derived from FHWA Table HF-10) ITEM 1995 2000 2005 User Fees (gas taxes, etc.) 84.1 100.6 114.6 Non-user taxes (gross subsidy) 21.4 29.0 39.2 Diverted to non-highway uses 21.6 16.6 21.4 Net subsidy -0.2 12.4 17.4 Total taxes used for highways 83.9 113.0 132.4 Net subsidy (% of total taxes) 0.0 11.0 13.1
Note that for 1990 (and earlier) the Table HF-10 doesn't exist. Table 88 90-11 (fix-me) (in "Highway Statistics" for 1990) is something like HF-10 but fails to show diversion of user revenue for other purposes. Per other tables for the States and the Highway Trust Fund, this amount seems to be $7.1 billion compared to 18.3 billion in non-user taxes resulting in a net subsidy of $11.3 billion But it's not clear if the "Imposts to Highway Users" includes or excludes diversions (the numbers don't "add up"). Nevertheless, it seems that net subsidy was over 10% both prior to and after the years around 1995. In Cox's table (see Discrepancy in Wendell Cox's Table the year 1994 shows a negative net subsidy of 7.4% (compared to negative 3.1% for 1995) while net subsidy is positive for both 1993 and 1996. Thus it seems that 1994 may have been the only year in which net subsidy was actually significantly negative.
HIGHWAY COST ALLOCATION STUDY, 1961 (87th Congress, House Doc. No. 54) Motor Vehicle Other Revenue Users Sources Interstate System 94.6% 5.4% Other primary federal-aid system 92.6% 7.4% Secondary federal-aid system 71.6% 28.4%
Note that only main highways are in the federal-aid system. So local roads are excluded from the above which likely had a larger share of financing by "Other Revenue Sources".
Note that in 1936 local roads costs (of counties and towns) were about 1/3 of total highway costs and about 50% of such costs were paid for by user fees. See Johnson p.552. For the 2/3 spent on main highways, it's not clear how much is subsidy since "Federal aid and grants" is not partitioned into user fees from federal taxes on fuel and tires and federal grants from the general fund (some of which was supposedly to provide employment during the great depression).
The many claims that highways are heavily subsidized are usually based on neglecting the diversions of highway user-revenue. In other words, they are looking at the total subsidy rather than the net subsidy. In one case (see NARP - Congressional Testimony it mentions the amount diverted for "Nonhighway Purposes" but neglects the amount diverted for mass transit which is listed separately in Table HF-10 but insidiously omitted from this testimony before congress.
Unfortunately, even government sites can be deceptive. The U.S.A. Census table only shows what's left of user fees after diversion of funds to non-highway purposes and titles it "Imposts on Highway Users". See Funding for Highways, by Level of Government statistics - USA Census numbers A small footnote mentions that this "excludes amounts later allocated for nonhighway purposes". In another case, data from the HF-10 table was used to create a smaller table with no mention of diversion of funds, but the amount of "highway user revenues" is shown net of the diversion to nonhighway uses with no mention of this fact on the webpage. See FundingForHighways EW-Gateway
There are various ways to estimate subsidy from Table HF-10. The question of net vs total subsidy has already been covered. At present, I think the most reasonable method is to just find the net subsidy: subtract the diversions to non-highway purposes (including mass-transportation) from the gross subsidy by non-user taxation. For details see Appendix: Calculation of Subsidy Using Table HF-10
One might alternatively ask the question: What percent of money spent on highways this year came from user charges. This method is flawed and double counts but it's the one used by Cox. Here's why it's flawed:
A significant amount of highway money is raised by bonds (debt financing). If the highway users will pay back the debt then one should seemingly count this borrowed amount as coming from highway users. But this is double counting since some of the highway user revenues are used each year to pay back debt and we would be counting these user revenues twice: once when the bond is issued and again when the highway users pay it back (in a future year of course).
Another reason it's flawed is similar. It's because some highway expenses are financed by investment income and drawdown of reserves obtained in past years when user revenues were higher than expenses and the resulting surplus was invested and put into reserves. This money of course came from users, but the money put into reserves was already accounted for in the year of the surplus. To again claim that this is money coming from users is double counting.
The same argument with a different twist applies to investment income. This is mostly not really income since it mostly serves to maintain the real value of the investment (if the interest rate is roughly equal to the inflation rate). Thus investment income can be treated to some degree just like reserves. To count it as a user revenue would be to some degree double counting (and more than double counting if the inflation rate for highway expenditures is greater than the interest rate).
Estimating subsidy by the method proposed in this report showed that for some years there was almost no net subsidy while for other years (including the 21st century) there is substantial subsidy. But one may question whether or not what has been called subsidy above is really subsidy as well as the alleged subsidy of heavy trucks by automobiles as discussed in the following sections.
Is what has been called subsidy to highways above always a subsidy? There is no definite answer to this question. Local streets and access roads have traditionally been paid for by local non-user taxes long before automobiles existed. One justification for paying for local roads by property taxes is alleged to be that land is of almost no utility unless there is access to it and that local roads are required for such access. This is the case for access both to city residences and rural residences where one needs to drive to town to purchase food and supplies. It also includes farm-to-market roads and the like.
Even though one doesn't use an automobile, access is still needed to maintain the property (deliver heavy furniture and appliances etc.) and also to supply building materials to construct buildings on the property.
An FHA report (see "Final Report of the Federal Highway Cost Allocation Study", U.S. DOT, FHA, 1982. p.I-5 and Appendix F) estimates that if non-user taxes are to pay for access costs, they should pay nearly 39% of road costs while they actually only paid about 25%. If one accepts these estimates, then the subsidy for road cost is flowing in the reverse direction: from the motorist to the non-user. The same report estimates that even if road users are assigned such access road cost, the non-users should still pay about 7% of road costs due to their benefits obtained from such items as street lighting, sidewalks, storm drains, etc.
Another point of view would argue as follows: Users of local residential streets (or local roads with low traffic) should be charged full user charges but since the gasoline tax is insufficient to cover these costs (due to low usage of such roads) the shortfall should be made up for by property taxes. Thus property tax revenues are supposedly only a temporary expedient until the practical technology is developed to charge motorists differential user charges (on different roads). On low traffic roads (such as in residential neighborhoods) motorists would then need to pay higher usage fees.
Another problem is that of separation of automobile and truck costs for their use of highways. It is often claimed that heavy trucks do the lion's share of damage to highways and are in part subsidized by automobiles. How much this amounts to is a debatable point, but even more difficult is the problem of allocating "residual" road costs between autos and trucks. "Residual" costs are the costs which the auto and truck share and are not attributable to the truck alone (for example the costs of excavation of earth when building the road, the costs for the thickness of pavement needed to only support automobiles, etc.). In economics, this is called "joint costs".
The method used for allocating these so called "residual" costs by the above mentioned FHA Cost Allocation Study is to charge the same for an auto-mile as for a truck-mile. This is likely biased in favor of trucks since a truck-mi would be expected to usually provide more utility than an auto-mile. This biased study then reports that only roughly 10% of auto user charges go to pay for road costs due to trucks (see p.I-13) but for an unbiased study the actual amount would be higher.
Results from a more recent study are summarized in Federal Highway Cost Allocation Study - Public Roads, January/February 1998 and Battelle Press Release: Highway Costs. url="http://www.fhwa.dot.gov/policy/hcas/addendum.htm" name="Addendum to the 1997 Federal Highway Cost Allocation Study Final Report"> The Addendum modified the results of the study and claims that combination trucks (the ones where a "tractor" pulls one, two, or three trailers behind it) only pay 80% of their allocated costs (with ones over 40 tons paying only 50% of allocated cost). The heavier (and larger) the truck, the greater the subsidy. Autos pay 100% of their costs while pickup trucks and vans pay 150% of their cost. Thus it appears that heavy trucks are being subsidized by the motorist and by smaller trucks.
The addendum shows makes it clear that the government considered the highway user fees diverted to the general fund for the purpose of deficit reduction to not be user charges. Thus when Congress stopped this diversion and the fees collected went to support highway costs, they considered that user fees from fuel had increased, resulting in combination trucks only paying 80% of their allocated costs instead of the previous 90%. But there really wasn't any change in user fees since motorists were still paying the same taxes on fuel and the subsidy to trucks really hadn't actually changed (if one considers the deficit-reduction tax a user fee -- this study didn't but the author of this report does).
The reason why pickups and vans pay more than their fair share is mainly due to the fact that they get worse miles per gallon and thus wind up paying more fuel taxes. This study assumes that such vehicles cost no more for highway costs than automobiles even though they weigh more and thus do slightly more damage to the pavement. But since heavy trucks are responsible for the lion's share of such damage the small additional damage done by pickups and vans is neglected. Thus all personal transportation vehicles that are "gas guzzlers" are paying significantly more than their fair share of highway costs and are subsidizing heavy trucks.
In most states, no sales tax is levied on motor fuel (gasoline and diesel). Such a tax exemption results in a subsidy to motor fuel and hence a subsidy to automobiles and trucks. Even when a sales tax is imposed, some of the money may be diverted to highway uses instead of going to the general fund. What justification is there for fuels to be exempt from sales taxes? Well, natural gas and fuels used to generate electricity aren't subject to sales taxes either (but perhaps they should be).
This failure-to-sales-tax-gasoline subsidy is large since fuel prices are high. The amount of sales taxes not paid would be perhaps roughly half of the amount of user taxes on fuel. But while this is a subsidy to motor vehicles, is it a subsidy to highways? Probably not, but it reveals a possible subsidy to highways explained next.
Recall that to find the net subsidy, we subtracted the amount of diversion of gasoline taxes, etc. for non-highway purposes. It turns out that a small part the money diverted is put into the general fund, just like sales tax revenue would be. Thus it could be considered as money diverted in lieu of sales taxes since its effect is just like a sales tax. Thus it would be viewed, not be a subsidy paid by motor vehicles for other purposes, but as a legitimate sales tax on fuel, which should not be deducted from gross subsidy. Even some of the money diverted to mass transit might be viewed in this way: it's like a special sales tax to finance mass transportation.
Just how much of the diversion is put into the general funds by the states? It's either a lot or a little depending on whether you look at gross or net. For 2005 (per "Highway Statistics" table DF) out of 63.54 billion of state user revenues, 2.11 billion (a lot) was put into the general funds. But as the same time, 1.85 billion was withdrawn from the general funds to finance highways. Subtracting this leaves only a net of 0.36 billion (a little) diverted from highway users into the general fund.
It's interesting that for some states, a large amount is diverted to the general fund while at the same time exactly the same amount is withdrawn from the general fund to support highways. The result is "no result" except for transactions on the books that cancel each other out. Why? It may be that the general fund is just a convenient fund to use for passing user fees thru to highway expenditures. It also allows flexibility in case the deposits from highway users and withdrawals for highway expenses become unequal it the future. But it also provides the opportunity for those who ignore diversions of highway user fees to claim that highways are subsidized from the general fund more than they actually are.
It's sometimes claimed that highway users should pay rent for the land on which the highways rest. If one takes the point of view that it's the highway users who own this land collectively, then they shouldn't need to pay rent for land that they already own. When the government builds highways, they often need to buy the land for the highway and the money to pay for this land comes mostly from highway users. For local streets, land for subdivisions (including land for streets) is purchased by developers and the cost of such land is included in the price of homes or lots. If the land was originally public land then one can alleged that the public land was given to highway users. So from this perspective, highway users have already paid for (or been otherwise granted) the highway land and shouldn't need to pay rent on it.
But in actual fact, the government owns the highway land, much of which was paid for by highway users. Should it charge rent for the use of such land? Public parks and forests generally don't charge any land rent to use them, although user fees are often imposed to pay for maintenance of them. But parks are a little different than highways since almost everyone can use parks while only people who own highway vehicles (such as autos, bicycles, etc.) can use highways.
From another perspective, land rents may represent unearned income and are immoral. Land was not created by people and thus land rent shouldn't be charged for its use, at least not by the government. One may claim that this is a Marxist argument, but its validity should not be judged by associating it with Marxism. From this perspective, the government has no business charging land rents.
So how much (if any) highways are being subsidized by failure to charge land rent for their use is an open question.
Even if one assumes that highway users own the highway system, the question remains as to why they are not paying property taxes on the highways. Does the lack of property taxes on highways represent a hidden subsidy to highways? One may claim that the value of access to property by streets and roads is incorporated in the price (and value) of taxable real property. For example, your home is more valuable since there is a street in front of it used to get to your home. Since property taxes are being paid on such property, then one may allege that part these taxes represent taxes on streets and roads that serve such property.
However, when private railroads were competing with highways for passengers, it seems to be inequitable that property taxes were being levied on railroad track but none on the highways. For freight, this inequitable situation persists to the present day. However, since freight railroads are inherently monopolies that are not well regulated today, it's not clear that reducing taxes on them will benefit the public.
The situation today, where highways are not heavily subsidized (if at all) is not true for bygone years. Prior to the introduction of the gasoline tax by various states during the 1920's, highways were heavily subsidized by the general taxpayer. Prior to 1920, only the modest registration fees paid by the motorist defrayed only a small fraction of the cost of the roads he used. But remember that prior to the 1920's there was no national network of paved highways and that railroad traffic grew as improved local roads provided improved access to railroads for passengers and freight. So in this way the subsidy to roads benefited railroads too.
Highways at first (before gasoline taxes) were mostly financed by local taxes, especially property taxes. But historically they were often built and maintained by "poll labor". For example in the 1820's the state of Indiana required that all able-bodied men work 2 days on the roads each year (without pay). However, one could pay money to avoid such duty. Poll labor was widely used throughout the world. Also, some road were built by the military and were available for public use.
In 1916 the federal government entered the picture with its "Federal Aid Road Act of 1916". It wasn't until 1956 that the federal government started to levy a user-tax on gasoline. Prior to this the federal government only levied only a 2 cent per gallon excise tax on gasoline, similar to the excise taxes it levied on other goods such as luggage, wallets and tires.
With gasoline taxes being enacted by states in the 1920s, the Federal Coordinator of Transportation (in in 1930s) concluded that since 1927, highway users were no longer being subsidized.
While prior to the late 1920s, highways were heavily subsidized by the general taxpayer, it's not clear to what extent (if any) they were subsidized after that. Frequent claims are made of high subsidy to highways but they tend to be biased due to their neglect of the diversion of highway user charges for non-highway uses as well as the benefits of streets to property owners. The purpose of much of such biased analysis seems to be to justify subsidy to public transportation.
All this doesn't mean that highways are good for the environment or that it was right to build them all and drive so much on them. Highways and the motor vehicles that run on them have very negatively impacted our environment, accelerated global warming, and due to importing of oil (and perhaps autos) contributed to huge foreign trade deficits for the U.S. But the fact remains that highways were for the most part built and maintained by fees imposed on their users or imposed on property owners who benefited from them.
Here are the details of one way to calculate the subsidy of highways using Table HF-10. The table appears in the annual publication "Highway Statistics" by the Federal Highway Administration (part of the U.S. Dept. of Transportation). Table HF-10 is titled: "Funding for Highways and Disposition of Highway-User Revenues, All Units of Government". It includes data from federal, state, and local governments.
The method is to first find the "gross subsidy" which is the
highway revenue from non-user charges such as property taxes,
income taxes, etc. Then the amount of money diverted from highway
revenue by mass transit and other non-highway purposes is found.
Subtracting this diversion from the "gross subsidy" gives the net
amount of net subsidy:
Net_subsidy = Gross_subsidy - Diversion
The amount of gross subsidy is the same as the "Subtotal" for "Other Taxes and Fees" shown near the middle of the table. This "Other Taxes and Fees" subtotal is the sum of 3 items in the table: 1. "Property Taxes and Assessments", 2."General Fund Appropriations", 3. "Other Taxes and Fees" (not otherwise itemized above).
The amount of diversion is found by adding three items from near the top of the table: 1. "Less: Amount for Nonhighway Purposes", 2. "Less: Amount for Mass Transportation", 3. "Less: Amount for Territories". The "Amount for Territories" is subsidy for transportation in U.S. territories (smalls islands such as American Samoa). Note that the "Amount for Nonhighway Purposes" mentioned above does not include the "Amount for Mass Transportation" since both amounts are subtracted in the HF-10 table from "Receipts Available for Distribution" to obtain "Net Used for Highway Purposes". Other items are also subtracted in HF-10 such as "Less: Amount for Collection Expenses" but this is not a diversion but is an expense chargeable to highways since it's a cost for collecting taxes for highways.
One might object to allocating all the collection expense to highways since some of what's collected is used for non-highway purposes such as mass transit. However, this oversight is approximately compensated for by neglecting to charge highways with the collection cost of non-highway revenues diverted to finance highways such as the cost of collecting property taxes.
Example for 1995 (in billions of $) using HF-10 (Revised Jan.
1997): Gross_subsidy = 21.4
Diversion = 15.8 (Non-Highway Purposes) + 5.6 (Mass Transportation) + 0.1 (for Territories) = 21.6
Net_subsidy = 21.4 - 21.6 = -0.2 Billion $. This is pretty insignificant as compared to the 92.5 billion $ of "Grand Total Disbursements" for 1995. Thus for 1995 there was essentially no net subsidy for highways: the taxes from highways users was just about equal to the expenditures for highways.
Note: If you haven't already, you'll need to read the above starting at Appendix: Calculation of Subsidy Using Table HF-10 to understand this section. The above method I used is somewhat similar to the method used by Wendell Cox to determine subsidy. See US Highways & Streets: Revenues & Expenditures. What I call the "net subsidy" he calls "BALANCING: Excess of Non-Highway Revenues over Expenditures". By "Expenditures" he means non-highway expenditures.s
For 1995 the value he shows for the "Excess" is -2.9 billion but I got only got -0.2 billion. Why the discrepancy? It's simply because he erroneously includes "Collection Expenses" under his "NON HIGHWAY EXPENDITURES" in his table. This is wrong since the expense to collect taxes paid by highway users (gasoline tax, license plate fees, etc.) is a legitimate highway expenditure in cases where gross tax revenues have been reported. Checking the numbers, the discrepancy of his -2.9 minus my -0.2 is -2.7 billion while the collection expenses that he failed to exclude were 3.0 billion (close to the 2.7 discrepancy). But there is still a residual discrepancy of 0.3 (not entirely due to rounding) which is due to my using the Jan. 1997 revision of the 1995 table while Wendell Cox likely used another revision (or the original issue) with slightly different numbers.
This discrepancy of erroneously including "Collection Expense" in Cox's "NON-HIGHWAY EXPENDITURES" was made for all years reported in his table (not just for 1995). I emailed him years ago about this and got a short response (that as I recall neither seemed to admit nor deny the error) but it hasn't been fixed yet.
What amounts to the same error is in Cox's table section "REVENUES COMPARED TO EXPENDITURES" subsection "Revenues over Expenditures" by which he means highway user revenues over highway expenditures. These values are identical to the ones in his "BALANCING" section that are due to including collection expense in non-highway expenditures.
But the errors in "Revenues over Expenditures" is calculated by subtracting what Cox considers to be "HIGHWAY USER REVENUES" from his "HIGHWAY EXPENDITURES".
One can use a search engine to find HF-10 but you will also find sites that just mention HF-10 like the site you are now reading. Or you can go to Highway Statistics and click on the year you want. Then click on "Highway Finance" and then find HF-10 and click on it. The problem is that if you want older ones you may need to download much more than just this one table.