Urban Form,Car Ownership and Public Policy: an Appraisal of Traffic in Towns

Traffic in Towns: a study of the long term problems of traffic in urban areas. Reports of the Steering Group and Working Group appointed by the Minister of Transport. London: H.M.S.O., 1963.

Its impact on government attitudes has been substantial, as witness the joint circular from the Ministry of Housing and Local Government and the Ministry of Transport to local planning authorities: 'The Government accept Buchanan's analysis of the situation.... They agree also with the main planning concepts and techniques set out in the Report.' Ministry of Housing and Local Government Circular No. 1/64, and Ministry of Transport Circular No. i, Roads, 1/64, p. 2.

'Traffic in Towns: the Critics Answered', Traffic Engineering and Control, Vol. 6, No. i, May 1964, pp. 40-41.

He estimates that above a critical size of town-approximately 80-90,000 for existing settlements-this demand will not be able to be satisfied.

'Can we afford Buchanan?', The Statist, 26th February 1964.

'... linear cities, annular cities, radial cities, satellite systems. It is a healthy sign that such ideas should be put forward, and the day may come when, suddenly, there is a crystallisation of views and it is clear that we should steer a new course. But this has not happened yet. For the period ahead into which we may reasonably peer-... a little beyond the end of the century-it is reasonable to assume that towns and cities will continue to exist broadly in their present form' (p. 30).

Hereafter references to Traffic in Towns are indicated by a page reference in brackets only.

It will be noticed that 'accessibility', as we have defined it, does not take the dual form it does in the Report. We simply regard vehicle users as wishing to minimise the disutility of getting from A to B. This formulation is more general and more useful, we believe. The Report's dual criterion is probably a reflection of the fact that the road investments for which it argues usually take the form of primary networks and distributors. Now, this form may well emerge from applying an investment criterion, but not necessarily so, and in principle one should define the benefits and costs from investment in a manner which is strictly independent of the form it may eventually take.

In reality, of course, things are more complicated than this simple division into two groups suggests. Thus, a resident in a given area may belong to both groups, so we should think in terms of roles rather than persons; however the broad distinction will suffice for present purposes.

Noise: Final Report, Cmnd. 2056, London: H.M.S.O., 1963, p. 3.

dBa is 'the unit of measurement of sound level' which allows for 'the varying sensibility of the human ear to different frequency'. ibid. pp. 154, 156.

For a method of measuring the cost in terms of 'accessibility', as defined here, of creating what is now termed an 'environmental area', see M.E. Beesley , 'Road Enquiry Deficiencies', Traffic Engineering and Control, Vol. 3, No. 11, March 1962-a discussion of the Oxford problem.

Though one might just be able to conceive of a study of what people are prepared to pay to see pleasant views, stately homes, or ancient monuments-distinguishing carefully how far in each case cars are allowed to intrude!

It may be asked-what would be the position of those who are renting or leasing property which is adversely affected by environmental changes? Now, as the value of property falls, the rents that can be secured also fall. The owners bear the loss: those who rent from them have the option of renting property elsewhere-they can maintain their standard of environment by moving-or they can take a cash saving. By contrast (long) leaseholders cannot so readily escape the loss, and would therefore be candidates for a share of the compensation.

P.J. McElhiney , in 'Evaluating Freeway Performance in Los Angeles ', Traffic Quarterly, July 1960, pp. 296-333, finds large decreases in the numbers of vehicles in streets parallel to new freeways-e.g. South Figueroa St carried 40,000 vehicles a day in 1955, before the Harbor Freeway was opened: in 1956, after it was opened, the figure was 14,000, in 1958 15,000. Similarly, in Chicago, F. F. Frye, in a careful before-and-after study of a 4-mile segment of the Congress Expressway, found a 7 per cent decrease in local streets in the 16 square mile area around it between 1959 and 1961, instead of an expected normal increase of 7 per cent. The decreases on paralleling arterials were much greater. 'The Effect of an Expressway on the Distribution of Traffic and Accidents', 42nd Annual Meetings of the Highway Research Board, Proceedings, Washington, D.C., January 1963.

At least rough justice would tend to be done in that the compensation on the one hand augments the road budget and makes more new work possible, and on the other that worse accessibility in particular areas raises the priority of road works in those areas.

We have not dealt explicitly with the important question of bringing time formally into the investment criterion. We could have done so, but the exposition would have been tedious and along well-established lines, and would in itself add little to our appraisal of the Report.

`Traffic-essential and optional: Essential traffic is the business, commercial and industrial traffic which is necessary to service and maintain the life of a community. Optional traffic is the traffic arising from the exercise of a choice to use a vehicle for a journey when the option existed either not to make the journey at all or to make it by some other kind of vehicle or form of transport. But traffic may be regarded as essential to the extent that buses are essential to carry loads which for various reasons cannot be discharged by individual cars. There is not necessarily always a very clear distinction between the two. Some apparently essential trips may, upon examination, prove less essential than some optional trips. In addition, if the distinction is made between the two at peak periods, some commercial trips could be called optional, in that they could be made at some other time of day.' Traffic in Towns, Glossary, p. 222.

The argument against subsidising public transport rather than increasing the prices in congested roads is that the former would under many circumstances lead to too large an expenditure and investment in urban transport as a whole.

From a study in progress of 1,500 Civil Servants' journeys-to-work .

Road Pricing: The Economic and Technical Possibilities, London: H.M.S.O., 1964 .

In fact the development of the land use traffic generation model perhaps may be conveniently dated from the Detroit Area Traffic Study carried out in 1952. For a discussion of the development and evolution of transportation studies in the U.S. see Walter Oi and Paul Shuldiner , An Analysis of Urban Travel Demands , Northwestern University Press , U.S.A., 1962.

For example, in discussing the number of zones used in the analysis (the number used for Leeds was ten employment and ten residence zones), they conclude that 'we could have obtained a finer picture had we been able to divide the city into a greater number of zones, but the number of cross-calculations that are involved increase rapidly as the number of zones is increased, and it would have been beyond our resources to deal with a much larger number' (p. 85). Existing computer programmes that can be used for making these assignments-they are being used in the London Traffic Survey for making assignments to over 900 origin and destination zones-are not mentioned. The planning department of the new town of Bracknell, with the Road Research Laboratory's advice, used such an assignment programme for an origin and destination matrix containing twenty-eight origin (residence areas in the Report) and twenty-eight destination zones. While the Ministry's study group might well plead having fewer resources than the London Traffic Survey, a claim of having less resources than the Bracknell planning department is less credible. It would seem that a study group commissioned by the Ministry of Transport ought to have at least as much access to the Road Research Laboratory's work as the planning department of a new town. (The first assignment for the existing town using the Road Research Laboratory's Pegasus II, a medium size computer, required 20 minutes of computer time. The second for the enlarged future town required 45 minutes. In addition punching of input tapes required 6 man days in each case-if professional tape punchers had been used instead of the authority's own inexperienced employees, the punching would have taken far less time. Of course, this does not include preparation of the zone-to-zone table for punching, and other clerical duties. However, the preparation of the zone-to-zone table was done by a commercial firm for £150, including punching of the 12,000 roadside interviews, processing, expansion of the sample to the actual volumes, conversion to p.c.u.'s, and conversion to 16-hour volumes.

As no information on this gravity model is presented in the Report, we cannot compare it to gravity models derived for use in other studies and in terms of the criteria suggested above for testing the model. It should be noted, however, that for any land use distribution, trip lengths and thus the amount of road capacity, depend entirely on the coefficients of this gravity model.

The assumption that the proportion of worktrips made during the peak hour will remain constant in the absence of restraint and with increased accessibility is virtually unjustified. The fact that as road capacity is increased, travel becomes more simultaneous has often been demonstrated. In the absence of pricing or some other form of rationing, removing urban congestion by increasing capacity is almost impossible, as increases in capacity only increase peaking. The Report treats the peak hour inconsistently in its four 'practical studies'. In Newbury, in contrast to Leeds, it states, 'it was apparent that the majority of the work-to-home journeys took place during the period from 5 p.m. to 6.p.m., with a distinct peak between 5.00 and 5.15 p.m.' (p. 55). In dealing with these differences, the Report states that, 'the duration of a peak period depends on a number of factors including the overlap that exists between the times at which firms commence or finish work, the relative location of homes and workplaces, and the existence of a state of congestion which may force or encourage lengthening of the period' (p. 62). In short, the Report, in treating the peak hour in each of the 'practical' studies differently, makes it unclear whether 'desired' or existing conditions should govern the degree of peaking the road system of a town should be designed for. 'As far as Newbury is concerned, we have assumed that the desire will remain for the great majority of all work journeys to take place within a period of one hour' (p. 62). The Report seems at various points to suggest that Londoners 'desire' a longer peak hour (2 hours) than the citizens of Leeds (i hour), and that both 'desire' a longer peak hour than those of Newbury (15 min.), and makes no effort to evaluate the effect of the proposed road improvements on peaking.

Of course, two or more positive and negative determinants might compensate one another. There is, however, no indication of such offsets to the Report's assumption of higher workplace and residence densities.

See, for example, Chicago Area Transportation Study, Vol. I, Survey Findings, December 1959, pp. 69-74, and Vol. II, Data Projections, July 1960, pp. 51-73, Wilbur Smith and Associates ; Future Highways and Urban Growth, February 1961, New Haven, Conn., pp. 68-95; J.F. Kain , 'A Contribution to the Urban Transportation Debate: An Econometric Model of Urban Residential and Travel Behaviour', The Review of Economics and Statistics , February 1964, pp. 55-64; J.F. Kain , 'Commuting and the Residential Decisions of Chicago and Detroit Central Business District Workers', Paper presented at the Conference on Transportation Economics sponsored by the National Bureau of Economic Research , April 26-27, 1963. (To be published in the National Bureau of Economic Research Conference Monograph Series), also available from the RAND Corporation , Santa Monica, Calif. , RAND Corporation Paper P-2735.

J.C. Tanner , 'Forecasts of Future Numbers of Vehicles in Great Britain', Roads and Road Construction, September 1962, p. 267.

The means and standard deviations of the sample are: A =282·4, SD of A = 59·2; D=8,578·3, SD of D=5,259·2; and Y=5,952-4, SD of Y=633-3. The Source of these statistics is the U.S. Bureau of Census: U.S. Census of Population: 1960, United States Summary, U.S. Government Printing Office, Washington, D.C., 1962.

We also wanted to be able to fit regression equations using net residential density rather than gross density. We did, in fact, fit these equations. However, since the fits were better for the gross density model (due, we believe, to inaccuracies in the residential land use variables attributable to widely differing survey dates-the surveys of land use used in calculating net residential density, and thus the net residential density variables, spanned a 13-year period from 1948 to 61), we used it to forecast car ownership. This land use data, as well as a discussion of changes in employment and population densities in these forty-eight cities may be found in John J. Niedercorn and Edward F. R. Hearle : Recent Land Use Trends in Forty-eight Large American Cities, The RAND Corp., Memorandum, RM-3664-FF, June 1962.

Central Statistical Office, National Income and Expenditure , 1963, London: H.M.S.O., 1963, Table 22, pp. 25 and 26. The median family income estimates using the internal purchasing power equivalents for mid-1960 are obtained as well as estimates using the official exchange rate as the latter is a notoriously bad deflator of the comparative purchasing power of national currencies. For a brief discussion of this question, as well as for the source of the internal purchasing power equivalent used here, see L. Needleman , 'The Burden of Taxation: An International Comparison', The National Institute Economic Review, No. 14, March 1961, p. 61.

To obtain comparability with the U.S. data from which Equation 2 is estimated, only land in urban use-i.e. excluding agriculture-was used in calculating 1960 population density.

U.S. Bureau of the Census : U.S. Census of Housing, 1960. Vol. I, States and Small Areas, Part I: United States Summary, U.S. Government Printing Office, Washington, D.C., 1963, p. XXXIX. The estimate of total cars per thousand of population is obtained by multiplying the estimates obtained from Equation 2 by i '089, an estimate of the difference in the two concepts. The value 1·089 was obtained from Chicago Area Transportation Study, Final Report, Vol. II, Data Projections, July 1960, Table 4, p. 13. While the data from which the coefficient was obtained refer to Chicago only, we would guess it to be fairly representative of cities of the size of Leeds and at minimum a better adjustment than none whatsoever.

Ministry of Transport , Road Motor Vehicles, 1961, London: H.M.S.O., 1962, Table 9, p. 9: Numbers of vehicles with current licences.

It should be noted that the selection of the sample used-i.e. U.S. cities similar in size and density to Leeds-already encompasses a large number of such factors. That is, Leeds may be thought of as belonging to the family of large cities, which differ in important respects in terms of the determinants of car ownership from small cities, towns, rural, and suburban areas. In this way the coefficients of Equation 2 already account for these differences.

We estimate that real median family income grew at a compound rate of 2 1/2 per cent between 1954 and 1962. In current pounds we estimate median family income in 1954-5 at £379, and in 1962-3 at £559. Adjusted for changes in the retail price level, the 1954-5 estimate is £458, in terms of 1962-3 pounds. These estimates were obtained from National Income and Expenditure, 1963, op. cit., pp. 25 and 26; Central Statistical Office, Annual Abstract of Statistics, No. 100, 1963, London: H.M.S.O., 1963, Table 361-Index of Retail Prices, p. 301.

'Traffic in Towns: the Critics Answered', op. cit., p. 41.

The concern shown by Leeds' planners about the capacity of the central industrial area after renewal took a very interesting form. During our visit they showed us, with some satisfaction and relief, the results of a recent industrial survey in which a fairly large number of existing firms indicated a willingness to occupy multi-storey or 'flatlet' industrial buildings provided by the local authority.

J.H. Niedercorn and J.F. Kain , 'An Econometric Model of Metropolitan Development', to be published in the Proceedings of the Regional Science Association, 1963 (also available as RAND Corporation Paper P-2663); 'Suburbanization of Employment and Population 1948-75 ', Highway Research Board, Proceedings of the 42nd Annual Meeting, 1963 (RAND Corporation Paper P-2641); idem., ' Changes in the Location of Food and General Merchandize Store Employment within Metropolitan Areas, 1948-58', Paper presented at the First Annual Meeting of the Western Section of the Regional Science Association, held in Berkeley, California, June 1962 ( RAND Corporation Paper P-2614-1).

According to information supplied by the Leeds City Planning Department, insured employment declined by just over 2 per cent between mid-1949 and mid-1961, while population increased by just over 1 per cent during the same period.

While few large towns could claim Leeds' reserve of open land within the city boundary and within the green-belt, nearly all have large adjoining tracts of land suitable for dispersal.

See Foster's estimate referred to above.

These cost estimates are obtained from: (i) Yk = Wc(O·311 +0·0708X)=0·086L, and (2) Yrow = Wroω(0·005X+0·00036X), where Yk = construction costs per mile in millions of dollars for a facility of L lanes, where X=net residential density of the area through which the facility passes measured in terms of thousands of population per square mile of land in residential use, where L is the number of lanes, where Wc is an index of the width of each facility as a ratio of the width of an 8-lane facility, where Wrow is an index of average right-of-way width, and where Yrow is the cost per mile in millions of dollars of right-of-way.

The details of the construction of these cost models and of the various tests of them are reported in a soon-to-be-published RAND Corporation Report: J.R. Meyer , J.F. Kain and M. Wohl , The Urban Transportation Problem. The cost models are derived principally from relationships found in Joseph Hyman, 'Costs of Constructing Transportation Facilities in the Chicago Area', CATS Research News, Vol. i, No. 14, pp. 8-12; Jones E. Watt, Jr. , and Marshall F. Reed, Jr. , Preliminary Cost Estimate of Construction and Right of Way for Rail and Highway Facilities in Alternative Regional Transportation Plans, Penn-Jersey Transportation Study, Paper No. 19 (mimeo); and from data obtained from a sample of construction and right-of-way costs assembled by the U.S. Bureau of Public Roads referring to segments of urban interstate highway in twenty-seven U.S. cities, and totalling over 550 miles in length.

This interpretation is shared by others far more skilled at evaluating Government attitudes than we are. The Guardian in a recent editorial on the Report says, 'From Ministers downwards it is officially and cordially approved. But how to attain this promised land? Is it a mirage? The Government has shelved it for the 1970's ... will Buchanan ever get off the ground? It looks unlikely, because it all seems so prodigiously expensive.' The Guardian ist Leader, Saturday 27th June 1964.