Lessons of congestion from the Goldman Sachs

A recent Business Insider article detailed how the Goldman Sachs managed to solve the long tolerated issue of lunchtime cafeteria crowding (unrelated humour-plenty Goldman Sachs video).  The analysts at the Goldman Sachs had identified that during the typical midday hunger-period of 11:30am to 1:30pm, the cafeteria was crowded, lines for food collection were long and stagnant, and for the average Goldmanite this was an inefficient use of their time.

The solution?

Apply a “timed discount” whereupon food purchases made before 11:30am or after 1:30pm receive a 25 per cent discount.  Or from the other side, discourage consumption during the “cost penalty window” where prices are inflated with a 33 per cent premium.

Any economics student could tell you that this is a simple problem of supply and demand, or more specifically a generic scarcity problem, where during certain periods of the day, there is a shortage of timely access to food (supply) and excess hungry employees (demand).  This shortage prompts lines! Lines mean waiting! Waiting equals underutilisation of labour! Oh no, the lost profits!

By decreasing the price outside the prescribed lunching time, Goldman has appealed to utility-maximising individuals with better-matched prices to their willingness-to-pay, or willingness-to-eat-at-a-non-traditional-meal-time, and hence succeeded in clearing the cafeteria market and eliminating this dead-weight-loss.

So what can we learn from the Goldman Sachs?

One situation – one which particularly frustrates me – is the market for road space.

Anyone knows that turning onto the Monash at 8am on a Monday morning is a bad idea.  Bumper to bumper. Start stop. Cars cutting each other off trying to find the elusive moving lane.  Forget the ludicrously low legal limit of 80kmh, VicRoads estimates that during the now even longer peak period of 6:30am to 9:30am, average speeds vary from 20kmh to 60kmh on our major motorways (link).

Congestion not only results in longer and more volatile travel times (which reduces our leisure, contributes to uncertainty, and interrupts schedules) but also leads to a higher probability and risk of road accidents; depreciation of vehicles through greater wear and tear; increased carbon emissions from longer drive times, lower speeds, and more frequent throttling; as well as stress, frustration and aggressive road rage.  From a fiscal perspective, the recent State of Australian Cities Report 2013 (link) estimates that congestion is currently costing capital cities $9 billion a year, with projections of $20 billion by 2020 and $6 billion for Melbourne alone.

So given the extensive economic, environmental, social, and fiscal costs and externalities of congestion, what is being done to solve this market failure of over-consumption?

At the moment, not much really.  Traffic is worsening and current policies are not working, nor is the debate improving (Fiona Scott anyone?).  Infrastructure and supply are crucial to addressing long-term congestion – improving public transport systems, removing railroad crossings, expanding road networks – but these things take years to lobby, years to agree, years to plan, and then they have to be built!

What really irks me is that if I’m picking up someone from the airport at 1am on a Sunday night when the freeway is ghostly empty, I’m charged the same toll or usage charge as someone jamming up the on-ramp at 8am on a Tuesday morning.  Shouldn’t the price of using the freeway and consuming our scarce road space reflect demand and supply, just as the prices of Goldman-approved sustenance reflect market conditions?  What are the incentives here, and what role do they play?

I know people who will catch the train an hour earlier just so that they don’t have to deal with the mass of commuters during the public transport rushes, and people who drive in early to avoid the gridlock of Melbourne’s peak hour roads.  But why not incentivise more people to do so?  What if the Punt Road exit was free if you beeped through before 6am? Or, what if it cost you double during peak hour(s)? Or triple? Or quadruple?

Cars already need e-tags, the scanning gates already exist, and prices already vary according to entry/exit points and type of vehicle. Why not add a time component to that as well?

In San Francisco, to cross the Oakland bridge during peak periods costs only US$2 more ($6 compared to $4), but since the introduction of temporal pricing, commuter delays have decreased by 15 per cent (link). The important realisation here is that cars using the freeway is not necessarily a bad thing, but rather it is the excess cars at the margin that breaches capacity and breeds congestion.  By varying prices even slightly, incentives can be adapted so as to shift marginal commuters away from peak periods, thus allowing for the free-flowing movement of the remaining drivers.

Smoothed or staggered price points would avoid choke-points and reduce volatility, whilst pre-setting prices would ensure consumer awareness and effective price information so that commuters would be able to adjust their driving habits ahead of time.  Another important pricing consideration is that due to the private-public partnership between The State of Victoria and Transurban, higher peak prices would be needed to offset cheaper non-peak prices and potential reduced usage, in order to compensate for any lost revenues and avoid litigation.

The pertinent point is that by appealing to people’s willingness-to-pay, price sensitivities and even social pressures – You drove in at 8am and paid what?! – market forces can help achieve an economically efficient outcome, whereby only those willing to bear higher tolls will consume road space during peak times, whilst those who are not will be pushed to off-peak times or onto public transport.

Concern may arise over a rebound or feedback type effect, whereby as a result of reduced congestion, non-drivers may begin driving and hence contribute back to the problem.  This would apply to individuals with a low tolerance for traffic and high price inelasticity, however, by definition their preferences would also ensure that the increase in road users would be limited to an upper bound where congestion is at a tipping point.

Another concern of course, is that those who are already departing early to avoid congestion may find themselves having to leave even earlier, but that may just be their utility-maximising outcome and the most society efficient outcome.  As well, road users may instead leak onto thoroughfares and arterial roads, but current drivers already do so, and most find it even worse and more costly than our blocked freeways.  Ultimately, these capacity issues further prompt the shift of commuters onto public transport, which is where investment in infrastructure and services is most needed.

So really, for now I’ve got one good solution for you – get a bike.  No traffic, no tolls (for now), no parking fees. Happy times …just hope it doesn’t rain.