Gas Energy etc Peter Berck 2003 Cost of

Gas, Energy, etc Peter Berck 2003

Cost of Cars • Private cost is mostly capital cost – Starts at 51 c/mile for a recent Taurus – Declines with age – IRS estimate of 37 c/mile (all ages) – Marginal costs about 13 c/mile

Taxes

External Costs

How to Regulate • Full marginal cost pricing – Hard to do, since must measure • Emissions • Dangerous driving • Place of driving • • • Gas tax (possibly by vehicle type) Reg. Fees Tolls (including for use of inner city) Emissions standards Refund/disposal fee systems for toxics: oil, batteries, tires • Subsidize alternative means of transport

Tolls • Proost and van Dender (1999) Brussels in the year 2005. – Cordon pricing results in about 50% of the welfare gain of full marginal cost pricing. • Geoghegan – Substantial gains from higher tolls on bay bridge, both congestion and pollution— pollution lowest at medium speeds.

Less Cars or Miles • Increased costs result – Less miles per car – Less cars – How much less?

Demand for Gas • Two Choice variables – Miles driven • Vacation choice (short run) • Bus/train v. car for commute (short run) • Location of job or house (long run) – Car fleet • About 10 years to adjust • Scrappage important • Many characteristics, including mpg, size, horsepower, safety, pollution

Rebound Effect • Demand Elasticity: – gasoline use (G), – vehicle miles traveled (VMT) – miles per gallon (MPG). • G=VMT/MPG. – VMT is a function of c, the price per mile of travel, which in turn – depends on the price p of fuel and the fuel economy, that is, – c = p / MPG. • • ln G = ln ( VMT( ln c ) ) - ln MPG ( p ), Elasticity formula. EG. p = EVMT. c ( 1 - EMPG. p ) - EMPG. Note that increasing MPG has indirect effect of making driving cheaper.

First Thoughts • Elasticity from derived demand – IRS estimate is about 37 cents per mile for full cost of driving – 1/25 gallon per mile at $2/gallon = 8 cents – A 1% increase in gasoline price is a –. 08/37 =. 0022= 0. 2% increase in cost of driving – Worse if one puts in time cost • E. g. 25 miles/hr at $10 per hour; 40 cents per mile! • Will take massive increase in price of gas to discourage gas consumption.

Another way • $cost of fuel price increase – – – 12, 000 miles per year 1/25 gallons / mile 480 gallons of gas per year 50 cent increase gallon (25%) = $240 additional expense • Or lower gas mileage – 1/12. 5 gallons per mile; 960 gallons; coincidentally $960 more – At 7% for 10 years: $7214 or about 1/7 the cost of a hummer.

Espey Meta Study

Short run elasticity • Short run keeps vehicle fleet constant – Regressions are quantity on current price, income, car fleet characteristics and size, etc. – Might include some lag structure

Long Run Espey

Long run • Long run allows car fleet to adjust – Appropriate price must be present value of expected gasoline prices. – Expected price changed much less than instantaneous price, so long run price elasticity would be woefully underestimated.

Real Gasoline Prices

$1, 132. 06 $1, 146. 62 $1, 145. 23 $1, 145. 96 $1, 153. 37 $1, 131. 05

Prices • • $1, 132. 06 $1, 146. 62 $1, 145. 23 $1, 145. 96 $1, 153. 37 $1, 131. 05 843 is 2002 value • Present value at 7% of previous 10 years of gas prices 1975 -1984; 1976 -1985; etc • Notice that pv price declines steeply post 80 -89. also that pv price is at top is 1. 36 times price at bottom—nothing like the 26 c to 2. 15 c per gallon nominal that I remember.

Espey meta results • Regresses (long run) elasticity on characteristics of study – Inclusion of mpg, car type, leads to less elastic estimate – Using only quarterly lags (1 qtr? ) leads to less elastic – 70% of response occurs within time frame of static models (nonsense. ) – Inclusion of countries other than US lead to greater elasticity (not surprising: $1 per gallon v. $4 gallon) – More long run elastic if estimated from more recent data.

Wheaton • Estimate, for countries by year: – Mpg – Autos per cap – Miles/ auto • As a function of gas prices and income. • Use three equations to get income and demand elasticity of gas.

Wheaton

More detail • Could track fleet: – New purchase decision • Exog: gas price, income distribution, topography, regulation, price of new auto(maybe) • Predetermined: existing auto • Endog: mpg, horsepower, longevity, number of autos, etc – Scrap decision • Same variables as above, but expect that income of lower quartile more important for scrapping – Also makes sense to track potential drivers.

Stoker Schmalensee • Income elasticity of demand – Key to forecast demand growth – Done partially non-parametrically • Adding drivers to equation costs effect of income on total gallons in half. – Inc elast for inc over 12 K is. 2 – Driver elast is. 6 – Implies that gallons will go up more slowly since number of drivers isn’t expanding quickly anymore. (not true for Calif or rest of

Price elasticity – RTECS doesn’t really have prices, just an average cost of gasoline by region. Hence estimate of price elasticity are spurious

Environment and Cars • Lit reviewed leads to strong link between gas price and mpg, however reducing fuel requires sustained high price. Doubling prices probably leads to 40% fuel savings. – Political suicide. – CAFÉ also saves fuel. No political will. – Tradeable mileage certificates would also do the job—need certifs to buy Hummer, get certifs if bought Prius. – Only way to get CO 2 down is to get fuel down.

Nox and Sox • Regulation directly makes cars cleaner (converters, in tune engines, etc. ) • Regulation adds to cost of car and decelerates scrapping, hence makes fleet dirtier on average from older cars. – Effect depends on how long cars are held when new car price goes up – Effect depends on how well pollution control equipment stands up on older cars. – Again, suicide to ban older cars, but ok to buy them up and scrap them.

Reality Check • Vehicle License Fee doomed Davis • VLF on a $10 K car was approx $200/year or $2800 present value (forever at 7%)

Welfare • IF V(gas price), welfare loss from taxing gas is standard. Makes no difference that response is to buy more efficient car. • If U(regulation) then welfare is much harder. Consumers don’t face a set of prices that would get them to elect the cars that regulators (and CAFÉ) chose for them.

Kling • Emissions are effluent standards in gms/mile. No trade or averaging. – Some vehichles beat standard so on he whole emissions are better than standard • Can trading do better?

Costs • To sell in CA, manufacturers must tell CARB what parts are used for pollution reduction. • Wang prices these parts to get cost of pollution reduction and that is the estimate of cost used by Kling – Could use ex ante engineering estimates, but these were too low. Or expert opinion. – Wang’s numbers are ex post

Regression • Cost is regressed upon MPG and emissions profile.

Least Cost • The cost functions are for each manufacturer and type of car (little, medium, big). Find min cost of current level of emissions • Minimize cost, holding types of car sold constant. • Allow car size to vary, hold consumer surplus constant. Greene has logit based measures of CS depending on car attributes.

Savings • On order of 10% for full trading.

Can One Emit Less? • Fullerton and West, 2002 – Tax cars for gas depending on type of car – Get very close to taxing NOx SOx etc. – Depends upon emission per mile being a function of car type and maybe age. – Amazing opportunity for fraud.

• Parry and Small (2001), welfare-maximizing gasoline taxes. – From a simple general-equilibrium model externalities also vary by mileage, whereas global warming varies by fuel use. – United States optimum is $1. 01 per gallon, acutual is 37¢/gallon – United Kingdom is $1. 34, less than half its current value. – optimum vehicle miles traveled (VMT) tax, 14¢/mile. – VMTs generate most of the external costs of driving, not fuel use.

Scrappage programs • Such scrappage programs have been examined by • Alberini et al. (1995, 1996), who found that the vehicles scrapped tend to be older and in worse • condition than other vehicles in that same age class. Although emissions from these vehicles are • high, their remaining useful life on the road is low. The cost-effectiveness of such policies ranges • from $3, 500 to $6, 500 per ton of HC removed, which makes them attractive in some contexts • and not others. Dill (2001) provides a summary of the studies on old-car scrap programs.