London has a problem with road closures. For example, New North Road has been partly or wholly blocked for several weeks by Thames Water. Thames Water works only one shift per weekday, so after 4 or 5pm and at weekends the site is empty and silent. Transport for London says that 17,200 vehicles run down New North Road on an average day. If we (conservatively) estimate five minutes of additional journey time for each vehicle when the road is closed, and price time at £20 per hour, we arrive at a per day economic cost of £28,000. That assumes one passenger per vehicle, excludes buses, doesn't count emissions or knock-on congestion elsewhere, and attaches no additional cost to unreliability. Now, it costs more to do work at weekends or evenings than in normal working hours. But assuming eight people working eight hours and normal wages for overtime civil engineering work, the *gross* cost of working evenings and weekends comes in at something like £2000-4000 per day –  and to be clear, that is *ignoring* the fact that working overtime reduces the work required at regular times. The net figure is probably £1000-2000 per day. So the costs of not working evenings and weekends are probably much more than 1000% greater than the benefits. The reason for Thames Water's sluggishness is, I gather, that it does not bear the costs that its roadworks create. Thames Water is behaving rationally given its incentives. But the system as a whole is wildly, incredibly irrational. London is burning money. This would not be not very hard to fix. Utilities companies should rent road space if they want to use it, and rents should be set at levels that incentivise behaviour that is optimal for society as a whole. Lane rental schemes have been shown to accelerate work and shorten closures, exactly as economic theory would predict. The Government wants easy, popular, pro-growth policies: surely there are worse places they could start.