Prior planning and preparation prevents… prices performing perfectly?

In its “Reformed National Pricing Delivery Plan and consultation on reforms to the siting and investment levers” document, DESNZ is consulting on locational charging, and on how those charges should interact with the Strategic Spatial Energy Plan (SSEP), the Centralised Strategic Network Plan (CSNP) and the connections process under Reformed National Pricing (RNP).  

This sounds very technical. But at the heart of the consultation is the interesting question of the purpose of locational charges in this new and more planned energy market context which we now face in GB. 

Formally, DESNZ’s consultation is about how to combine what they are calling siting and investment levers. Put simply, these are the mechanisms government uses to influence where projects get built: some create the underlying conditions for development (“Enabling levers”), while others directly steer or constrain investor decisions (“Primary levers”). Of the primary levers, the government is considering what job locational charges should now be asked to do. 

Historically, the answer was fairly clear. Locational charges were meant to send signals, however imperfectly, about the network costs users imposed in different places. Developers should face those costs, internalise them, and choose locations accordingly. In that world, locational charges were a tool for encouraging efficient siting in a competitive market. 

But the advent of SSEP changes that context. If strategic planning is now meant to determine that the country wants X GW of technology Y in location Z, and if connection “queues” are being increasingly curated, then maybe the role of locational prices changes. So the consultation can be read as raising a bigger question: are locational charges still mainly supposed to be reflective of network costs, or are they increasingly being asked to help reproduce a centrally determined spatial plan?  

The first step to answering that question might be to answer a prior question: is “hitting the SSEP” a good policy goal?  

There may be a very good reason not to try to hit the SSEP exactly. The SSEP is a plan. It will presumably be developed using the best information NESO and industry can assemble. But it will still be imperfect. Technology costs and productivity will move. Demand may emerge differently from expected. Network delivery may lag the plan. And the plan is only intended to be updated periodically. So at any point in time, the efficient pattern of generation and storage may not be exactly the one the planners produced (and this may be true even if the networks are on track to hit the CSNP). 

If that is right, “traditional” locational signals may still have an important role to play, helping the market internalise the network costs of locating in different places, and allowing investment to deviate from the plan where that is efficient. 

There is one important change from the old world, however. In a strategic-planning context, cost-reflective cannot just mean reflective of today's network (or indeed, ignoring any information about actual or planned future capacity). It has to mean reflective of the network as planned. If cost-reflective locational charges ignore the CSNP and are based only on the current system, investors will not properly take account of the network that is intended to exist. That would risk sending the wrong signal altogether. 

But what if policymakers really do want to hit the SSEP? If charges are calibrated to be reflective of network costs, you may not get the spatial plan outcomes (because the market’s assumptions on things like cost and productivity of different sites might not be the same as those of the planner). So the way you think about what determines the “right” locational signals has to change. But it might also be reasonable to ask whether locational signals are the right tool to use in this situation. 

The direct way to achieve the spatial plan, at least for supported technologies (which today is practically all of them, if the capacity mechanism is included as “support”) is to allocate technology-location combinations through the connections process or through allocation rounds. 

Doing this via the connections process would involve ensuring that Gate 2 offers were only awarded to projects which strictly fit within the SSEP constraints. This might leave some locations with too few projects in some technologies, and might also mean that allocation rounds become materially less competitive. This is because the project pipeline may be no larger than the quantity being procured, weakening competitive pressure and increasing the scope for developer rents. This is true even if allocation rounds are spread out, as in earlier rounds developers will anticipate the potential for scarcity in later rounds. 

The alternative is for the connections process to be more permissive (i.e. materially more Gate 2 offers are made for each location and technology than required by the SSEP), with the allocation process determining which projects are built. This could be achieved by increasing the locational granularity of the auctions (currently the only locational differentiation in the CfD allocation rounds is Scotland vs. England and Wales). However, this would be a significant change, so the alternative would be to retain national allocation rounds and use locational charges to achieve the SSEP outcomes. 

It is not clear how easy that would be. Let’s imagine we stick with today’s national allocation rounds and that we also stick with locational charges under which all plant of a given technology and in a given location face the same charge (irrespective of capacity invested).  

These charges are too blunt to deliver the required SSEP outcomes. They can increase the costs (and so allocation round bids) of all plants in a region by a broadly similar amount, but they cannot differentiate in a precise way between projects with the same technology. Using them to achieve a location-technology combination will be difficult. 

This might lead you to think about locational charges which vary with the amount of capacity allocated. Under such a regime, the more capacity that is allocated support in a given location (and hence the closer to the relevant SSEP limit we get), the higher the charges would go. The aim would be that, at the upper SSEP constraint in each location, the charges are so high that no further capacity is successful in the allocation round (in locations where there is a risk of not achieving the lower SSEP constraint, the charges would instead reduce and presumably possibly go negative).  

This regime might get closer to the SSEP outcomes (albeit in a way which may be complicated to calculate and controversial to apply!). But is this regime superior to a series of more location-specific auctions? Some might say “yes”, because you achieve more competitive tension. A national allocation round auction sounds more competitive than a series of more location-specific auctions. Right? 

Not necessarily. If locational charges vary by capacity as described above, the projects which are choked off by the higher charges arguably cease to provide much of a competitive constraint (their bids are forced to be high by the charges, and bidders with lower cost projects will expect this). It might look as though projects across all regions are competing in one large auction, but in substance they may not be. It might amount to a national auction in name only. The risk is that this form of locational charges is simply a more opaque way of implementing the same regional allocation as you would get through more locationally specific auctions. 

So the real question underlying this consultation is really the job locational charges are being asked to do: 

• if government wants the market to have room to discover better outcomes around a centrally determined spatial plan, then locational charges remain useful, provided they are genuine cost-reflective signals grounded in the future network the CSNP is meant to deliver; but 

• if government wants the SSEP to dictate location, then locational charges look increasingly redundant. The type of locational signals needed to deliver the government’s objective may be very different to those we see today (in particular they may need to vary to reflect how close we are for a given technology and location to the SSEP constraints), and using administrative prices to engineer a planned outcome looks unnecessarily indirect. 

And as a parting thought, it is also worth noting that the government’s objective may not be uniform: it may be both more keen and more able to hit the SSEP in some technologies than others. In which case, that may point to very different approaches by technology. For example, this might involve a more plan led regime for some (large, locational critical?) technologies, and a more cost-reflective locational charging regime for others where efficient deviation from the plan matters more.