Beyond a curmudgeonly few, there is little debate now on the efficiency case for levying user charges. Harry Clarke

Transcription

1 Beyond a curmudgeonly few, there is little debate now on the efficiency case for levying user charges. Harry Clarke 1

2 Congestion charging: a curmudgeon s view Mark Harrison Roads, cars and taxes Crawford School of Public Policy Seminar September 9,

3 Who could be against a Pigou tax? even though economists were able to provide a perfect blueprint for beneficial State action, politicians are not philosopher kings and a blueprint might quickly yield place on their desks to the propaganda of competing pressure groups. Fancy finance, like a fancy franchise, whatever its theoretical attractions, has, at all events in a democracy, dim practical prospects. It is not sufficient to contrast the imperfect adjustments of unfettered private enterprise with the best adjustment that economists in their studies can imagine. For we cannot expect that any public authority will attain, or will even wholeheartedly seek, that ideal. Such authorities are liable alike to ignorance, to sectional pressure and to personal corruption by private interest. A loud-voiced part of their constituents, if organized for votes, may easily outweigh the whole. Arthur Cecil Pigou 3

4 We need to evaluate actual projects and reforms on whether they improve efficiency Do governments have the information or incentive to implement an efficiency raising scheme? Congestion charging is likely to be inequitable so it had better be efficient. Yet the government could easily introduce an inefficient scheme. Need to worry about what is done with the revenue. The standard avoidable cost of congestion figures exaggerate the potential gains from road charging and provide little guidance for sound policy development. So called double dividend argument for taxing externalities

5 Basic model of highway congestion Assumes all drivers have an identical cost of time $ per trip D SMC AC C 0 X 1 X 0 Traffic Flow 5 (completed trips)

6 Measures of the costs of congestion Texas Transportation Institute measures relative to free flow level (red area). $ per trip D SMC AC C 0 X 1 X 0 Traffic Flow 6 (completed trips)

7 Measures of the costs of congestion BTRE measures deadweight loss (purple area, only a fraction of the red area). $ per trip D SMC AC C 0 X 1 X 0 Traffic Flow 7 (completed trips)

8 $ per trip A congestion charge can achieve the efficient outcome. The efficient charge is the marginal external cost at the optimum (X 1 ) (internalizing the externality). D SMC AC + T* AC Toll X 1 X 0 Traffic Flow (completed trips)

9 $ per trip Those tolled off lose the blue area Continuing motorists lose green area (pay red plus green in tax, reduction in congestion costs to continuing motorists is red area). Motorists lose green plus blue area. D SMC AC + T* AC Toll X 1 X 0 Traffic Flow (completed trips)

10 Some implications for congestion pricing Ideally, charges should reflect marginal external cost at the optimum, which is less than currently observed levels Charges should vary with external costs (by time of day, day of week) Congestion is not eliminated. Part of the policy challenge is to determine the efficient level of congestion never zero Drivers lose which is why it is so politically difficult to introduce road charging. The welfare gain is small relative to the loss in consumer surplus and the gain in revenues. If some of those revenues are wasted, then it takes proportionately very little of it to eliminate any efficiency gains from congestion charging. Charges can be too high we can have too little congestion. For example, monopoly pricing

11 $ per trip D SMC AC Toll X 2 X 0 Traffic flow Congestion can be reduced too much: possible efficiency loss from overcharging is the red triangle, which may be bigger or smaller than the purple one from excessive use in the open access regime.

12 Complications Different types of congestion: hypercongestion, bottlenecks. Costs of congestion are greater than extra travel costs: schedule delay, uncertainty. Heterogeneity Value of time differs between drivers and according to circumstances Rationing by queuing is inefficient These possibilities mean larger gains from congestion pricing, drivers as a whole may gain.

13 Estimates of the gain from congestion charging The BTRE avoidable cost of congestion (purple area) indicates the potential gains (if the model s assumptions hold true) from a costless, perfect pricing scheme, which sets price equal to marginal cost at all times and results in the efficient level of road use. Magic wand economics. In practice, actual congestion pricing schemes are expensive, imperfect, and not able to capture the theoretical benefits identified they do not involve perfect marginal cost pricing. The gain will be much less. Imperfect schemes: partial charging, cordon schemes, charges that don t vary by time of day. Proliferation of exemptions that serve to undermine scheme objectives Partial charging increases congestion on uncharged roads, net effect may be to reduce efficiency. For example: tollroads, HOT lanes. 13

14 Road pricing schemes are costly to set up and operate. The London Congestion Charge required start up costs of 200 million in 2003 pounds. The 693 fixed cameras cost more than 100 million. Operating costs are about half of revenue. Singapore: The cost of operating the scheme has remained about per cent of total revenue collected The total start up cost of the Stockholm trial was 1.9 billion Swedish Kronor in 2006, around $A320 million at current exchange rates. Operating costs per cent of revenue. In the Netherlands the cost of implementing a national system of congestion charging was estimated in 2006 as billion (2006 euros) with operational costs estimated to be million per annum.

15 Different methods of road pricing and international experience Cordon schemes: tend to be small: Singapore, 7.25 km 2 ; London, original charging area: 21 km 2 ; Stockholm, 23 km 2 Once within the cordon, there is no discouragement to unlimited travel, may generate congestion at boundary points. Overseas schemes have been successful in reducing congestion, but have been confined to small areas, in cities where congestion is severe, and where an effective public transport network provides options and responsive demand. And yet are still rarely justified in terms of net community benefit. In Australian cities, demand for car trips is likely to be less elastic and the charges may induce little behavioural change.

16 $ per user D SMC C 3 C 2 AC Toll C 0 C 1 X 1 X 0 Number of vehicles When demand is more inelastic (lower responsiveness) Revenue increases (red plus green) Loss to motorists increases (green area) 16

17 $ per user D SMC C 3 C 2 AC Toll C 0 C 1 X 1 Number of vehicles Increases opposition Potential efficiency gains fall (compare blue and purple triangles) 17 X 0

18 With a less elastic demand, the chance of an inefficient outcome is greater. Revenue is high, especially relative to the low potential efficiency gains. It is more likely that congestion will be reduced too much. It doesn t take much for implementation or revenue collection costs to swamp any welfare gains. The diagram assumes revenues transferred to the government are a mere transfer. But if some of the revenue is wasted, such as being spent on projects with costs greater than benefits, even a small proportion of waste could outweigh any efficiency gains from congestion charging. In particular it shouldn t be automatically assumed that the proceeds should be spent on public transport, as is done in most overseas schemes. 18

19 The avoidable cost estimates do not measure the benefits from other policies to alleviate congestion Environmentalists claim Building more roads to ease traffic is kind of like trying to cure obesity by loosening the belt the avoidable costs of urban congestion may grow to around $20 billion in This cannot be reduced simply by building more city infrastructure, as most new road space induces new traffic. (Henry Tax Review 2010, p.53). The net benefits from increases in road capacity and subsidies to public transport depend on the change in total congestion costs (which are real costs borne by drivers) not just the change in the deadweight loss triangle (excessive costs). For example, a beneficial increase in road capacity may increase the BTRE s measure of congestion costs.

20 Benefit from increased road capacity is shaded area. This can be positive and B > A. MC 0 AC 0 MC 1 AC 1 AC 0 AC 1 A B D C X 0 X 1 20

21 Interactions with existing taxes and subsidies Say we put on a congestion charge and this increases public transport demand and reduces petrol demand. This gives the shaded loss in the related markets, reducing the gain from congestion charging. Alternatively, a congestion charge reduces the optimal petrol tax and public transport subsidy. P X D 0 D 1 P Y D 1 D 0 MC MC Y +T Y MC subsidy MC Y X 0 X 1 Public transport trips Y 1 Y 0 Petrol (L)

22 P X Is there a double dividend? The double dividend argument: a gain in from reducing excessive congestion and a further gain from using the revenue to reduce distorting taxes (revenue effect). The largest distortion is in the labour market: income taxes, consumption taxes, benefit phase outs. This claim ignores the tax interaction effect. Say the congestion tax reduces labour supply. S 1 S 0 W (1 t)w H 1 H 0 Hours worked 22

23 Here the congestion charge pollution tax exacerbates the labour market distortion. The relevant issue is what happens to government revenue in the compensated equilibrium Marginal deadweight loss of taxation (MDLT) is the efficiency cost associated with raising an extra dollar of revenue. If we were at the second best optimum, where the MDLT is equated across taxes, the tax interaction effect offsets the revenue effect and there is no double dividend. If you don t start at the optimum, the tax interaction effect may be bigger or smaller than the revenue effect. These tax interaction effects can be significant 23

24 Parry & Bento empirical study: a congestion charge reduces labour supply and the resulting welfare loss is double the gain from reduced congestion. But if the revenue is used to reduce labour taxes, the overall welfare gain from congestion charges doubles. Say gain from road charging is 10 cents for an extra dollar of congestion charges. And the tax interaction effect is a 20 cent cost. Then the MDLT for the congestion charge is 10 cents. But if the MDLT for the labour tax is 30 cents, using the revenue to reduce labour market taxes gives a net gain of 20 cents. Not a strong case for congestion charging But better to use for this purpose than some of the others. But will congestion charging reduce labour supply? 24

25 Effect of congestion charging on the labour market We use the standard labour leisure model. The consumer has an endowment of time, T hours and an endowment of non-wage income $A. The fixed endowment of time must be allocated between different uses. We assume just three uses: L is hours of leisure. Leisure is what we call all non-market activities (time spent not working or commuting). F is the fixed time cost of going to work (hours spent commuting). We assume time spent travelling to work is valued the same as time spent working. 25

26 Congestion charge: increase in money cost, reduction in time cost, total commuting cost increases $C per day C 0 C 1 A M 0 A M 1 E=(T,A) (T, A M 0 ) (T, A M 1 ) L 1 L 0 T F 0 T F 1 T L hours per day 26

27 Increase in money cost, reduction in time cost, total commuting cost increases Item Substitution effect Income effect Total effect Leisure Hours worked + + Consumption Wage income + + The change lowers real income, which reduces L and C. By the time budget constraint, T = H + L + F. As F and L fall, H rises and so wage income rises.

28 Worker drops out of labour force $C per day C 0 C 1 A E=(T,A) A M 0 A M 1 L 1 L 0 T F 0 T F 1 T L hours per day 28

29 The effects of congestion charging on the labour supply So on the intensive margin, congestion charging increases H. Those who still work, work more. On the extensive margin it reduces H (some drop out of the labour force) Parry and Bento assume the only response is on the extensive margin. 29

30 Conclusions The need for case by case evaluation (cost benefit analysis) The relevant policy issue is the costs and benefits of feasible policy responses and which has the greatest net benefit. Justifying a road pricing scheme requires rigorous bottom up, case-by-case analysis using network models and tailormade strategies, location and project specific. The key issue is not the estimated cost of congestion, but the returns to the options for reducing congestion. That requires detailed network analysis and calculations regarding the impact of the policy proposals. The optimal charging system would vary from city to city, determined by the topography of the city, the shape of the road network and the nature of the traffic flows.

31 Without efficiency-based cost-benefit analysis of all policies to deal with road congestion, including use of revenue raised, governments run the risk of lowering social welfare. For example, without rigorous project appraisal, the political process could produce a congestion charging scheme which is both inequitable and inefficient Conclusions on pilot programmes Should be targeted: Where congestion is high. Where demand is elastic. Where value of time is the highest and where there is heterogeneity. Partial coverage means need to worry about scope for diversion.