April 2005, updated 6 October 2005
John F. Raffensperger,
Dept. of Management, Univ. of Canterbury, Christchurch, NZ.
Phone 03-364-2987x8616.
Mark Milke,
Dept. of Civil Engineering, Univ. of Canterbury,
Christchurch, NZ. Phone 03-364-2987x6248.
A grant from the New Zealand Hydrological Society (NZHS) allowed the initial development of software for the on-line auction of groundwater. The system under development at the University of Canterbury uses standard hydrology software (MODFLOW) and open source linear programming code. To our knowledge, this is the first general model for a fresh water market in the world. The grant went towards the employment of a part-time programmer, Kyeong Yoo, to work with the researchers during the summer of 2004/2005. The support led to the demonstration of the software to a number of interested groundwater parties in the South Island, including a demonstration on 20 January at the University of Canterbury with about 10 interested parties including consultants, regional council personnel, researchers and farmers. As a result of the demonstration, a few further refinements to the software were made until the funding had been used completely in mid-February 2005.
The funding allowed substantial work on software development that would otherwise still not be completed, delaying the advance of this bold new technology. The software is linked to its own web server and so is not easily downloaded. Please contact John Raffensperger to see a demonstration of the University of Canterbury effort. Because of the availability of a high-quality software programmer, it was decided to focus our efforts on software this summer, rather than to also develop information on supply/demand curves, as was proposed in the grant application. This report summarises the work conducted under the Grant and also gives an overview of the project itself.
The goals of water allocation are to first steward the environment for future generations, and second, above the environmental bottom lines, to maximise the value of the resource to New Zealand in an equitable way. Decades of economic research have shown that markets allocate resources better than quotas, government directives, or legal decisions. Markets maximise benefits to society, if the market is properly implemented and regulated. Where users can be made better off by trading, trading should be possible. Without individual incentives to improve water management, users will waste water and damage the environment. Without trading, in the long run, New Zealand is likely to have worsening water shortages.
The problem to date has been that no one has worked out how to implement a market for water. Economists stress that a market is the best way to allocate resources, but they do not describe in detail how to set up the market. A very few researchers and organisations have developed trading systems, but those seem to have shortcomings. Consequently, water markets set up in other places have resulted in numerous problems, including unclear property rights and damage to the environment. As far as we can tell, no market for fresh water exists anywhere in the world that guarantees protection for the environment, protects third parties from the behaviour of traders, and gives correct price signals to users.
New Zealand fresh water is a public good. Water must be viewed as a common, shared resource. This implies that all ground water and water in public waterways must be owned in the long run by government. Furthermore, government cannot give permanent title to water because the availability of water is variable, and because all users draw water from the same resource. Therefore the use of a market for groundwater is not the same as the “privatisation” of water. Trading privately owned water will not work like trading kettles and kayaks, because a trade of water between two users can injure a third party or the environment.
New Zealand has led the way for sustainability through development of innovative markets. In 1986, New Zealand became the first country to use quotas nationally for multiple fish species. Ten years later, New Zealand developed the first modern trading system for electricity. While not without controversy, these systems have proven very good for New Zealand and the systems are now emulated around the world. We think that New Zealand could again lead the way.
Following the new technologies for markets, especially in energy, the University of Canterbury is developing new ability to model water systems and to manage bids. The technology allows users to buy and sell water conveniently, while respecting environmental standards. The technology has the potential to improve protection of the environment, improve New Zealand ’s economy through better use of water, and simplify the resource consent process. To bring about these improvements, it will be important to move from the technology from a University incubator into the real world. Before describing the software developments made possible by the grant from the New Zealand Hydrological Society, it will be helpful to see how we currently envision how government could task a market manager with fairly allocating water, and with protecting the environment, using this technology.
A market manager acts as a broker, under the regulation of the regional councils. Each catchment would have its own market manager, because the catchment areas are, hydrologically, relatively independent. The market manager is obligated to respect the environmental standards as set by the community, and within those standards, to allocate water in the best economic manner. The market manager could be the council, a state-run enterprise, or a co-operative of users.
The market manager will operate a standard hydrology model, a web system to accept users’ bids, and an operations research program for allocating water and setting bids.
In our system, a hydrology simulation calculates the effects of each well on its neighbours and the environment.
If well 1 takes out a unit of
water, how does this affect all other wells and the required environmental flows?
If well 2 takes out a unit of water, how does this affect all other wells and the required environmental flows?
The Forever Fair system currently calculates the effects for each well on every other well using the program MODMAN (http://www.geotransinc.com/modman.html).
The market manager is required to enter required standards for the environment, as set by the government and the community, assisted by hydrologists. Examples include head difference constraints to maintain flows in certain directions, draw down constraints to protect the aquifer, minimum flows in surface water bodies, and several other types. These standards ensure that the environment is protected over time, and these standards can be adjusted easily at any time, as needed for the environment. This is the “forever” part.
Users bid through a web page. Bids are changes to their existing allocations, selling (minus) or buying (plus). To sell water, a user can name a price at which he is willing to sell. If the market accepts his bid, he is guaranteed to get the price he asked or a higher price. If he wants more water, he can name the price at which he is willing to buy. If the market accepts his bid, he is guaranteed to get the price he asked or a lower price.
To understand bidding in the Forever Fair System, we need to understand “value for the next megalitre,” which is the user’s marginal value for water. This is the concept that the first megalitre of water does the most good, and the last megalitre of water does the least good.
1st megalitre. Replace recently lost moisture. $3
2nd megalitre. Add extra moisture for today. $2
3rd megalitre. Give a good soaking. $1.5.
4th to 5th megalitres. Give a deep soaking. $1
6th to 15th megalitres. Heavy soak with run-off. $0.50
Here is another way to think about marginal value: given a list of prices, how much water would you buy at each price?
If the price were $3/ML, I would buy 1 megalitres.
If the price were $2/ML, I would buy 2 megalitres.
If the price were $1.50/ML, I would buy 3 megalitres.
If the price were $1/ML, I would buy up to 5 megalitres.
If the price were $0.50/ML, I would buy up to 15 megalitres.
Of course, a user will probably want to sell at a high price and buy at a low price. The point is that for any trade, users have different values for different quantities of water. The bidding system is designed to capture these different values.
The system runs several tentative auctions, with only the final auction being firm. In this way, users can try out bids to see what will happen, and have a chance to adjust their bids before they commit.
Auctions would probably run about once per week, more often in drought. The frequency depends primarily on the relative transaction costs. (If operating the auction is expensive, it should run relatively less often.) Users can enter bids for future auctions. If users do not bid, a computerised system can allow automatic re-use of the previous bids. Therefore, users do not have to be handcuffed to the bidding screen.
After the hydrology simulation and bidding, the Forever Fair system reads the bids as entered by the users. The web system connects to a database and standard hydrology software, with quantitative environmental standards. Based on the bids, hydrology, and environmental standards, the system allocates water to users and sets a price at each well, in a way that respects the environment first, and then maximises the public good within those standards. The allocation and price setting are done with a standard operations research model.
The operations research model calculates the flows that will make everyone as happy as possible, within the environmental standards. The system will reallocate water from a user who will sell at a low price to a user who will buy at a high price, adjusted by the hydrology. The final quantities of water make every user better off compared to no trading.
The model calculates prices as the marginal cost to the economy (the lost benefit to all farmers) if the well takes another unit of water. Prices depend on the users’ bids, the hydrology, and environmental protections. Therefore, different wells have different prices, even if all users bid identically, because the hydrology and the environmental sensitivity is different throughout the catchment. These underlying mathematical and economic principles are the same as those of the NZ electricity market. The Forever Fair system sets prices so buyers buy at or below their reservation price and sellers sell at or above their reservation price. This is the “fair” part.
Once a trading system is in place, New Zealand would have new options for simplifying the resource consent process.
Using the existing consent process with the Forever Fair trading system, consents could remain as they are. However, there are problems with this. Because those with water permits can sell if the price is right, there would be dramatic increases in requests for permits in underallocated aquifers. Those first to the trough would have an unfair windfall (especially if they always sell), and new users would have no convenient way to enter.
Another way that users could benefit at the expense of the market manager is ff users were given overly generous resource consents, so they have legal permission to use more water than is actually available. Then the computer program would require the market manager to buy water from the users in order to protect the environment. We can show mathematically that users could bid up prices to infinity, thereby improperly profiting from environmental protection at the expense of the rest of the country. We do not see this as tenable.
On the other hand, with a pure user-pays scheme, there would be no resource consents or permits and users must bid every time they wish to use water. We can see social and political objections to this system. We can show mathematically that a user-pays system is not necessary for an economically efficient allocation of water.
We have examined the middle ground (where users trade) between the extremes of “user pays” and “user wrongly profits”. To do this, the Forever Fair system adjusts short-term consents proportionally for all users in such a way that total revenue paid to the market manager is zero (except a fee for operating the trading system and for related services such as hydrology measurements). Even in drought conditions, some users would pay and other users gain, but the net balances to zero. Users pay each other, not government. The prices and allocations are the same as under a user pays system, but because the initial consents have been adjusted, each user buys or sells from a different initial wealth, and always better than where they began. The result is a pure user trades system, which takes into account the variable amount of water available at every auction, and the required flows for the environment.
This user trades system then allows a greatly simplified resource consent process. For allocating new resource consents, and for downward adjustment of unused consents, we propose consideration of a simple “buy-up/sell-down” process. Under a “buy-up/sell down” process, users’ consents would revert to the water used last year. Each year, the market manager would give each user a new consent equal to the amount actually used in the previous year, net of trades. This follows the “use it or lose it” principle. In the new year, each user can then bid to sell from this consent or bid to buy in addition to it, subject to the hydrology and the environmental standards. If the user neither buys nor sells water, the user neither loses nor gains money, and the user retains the consent. Similar schemes can be devised where a certain fraction of the unused or newly used water is subtracted to or added to the consent, though this can lead to difficulties.
With a “buy up/sell down” scheme, a new user would first get council approval for the type of use. The new consent process would mainly be to confirm that the water use is legal, to ensure proper drilling techniques are used, and to provide information for a register of permits (e.g., location of water take). For example, a user who wants to grow Brussels sprouts would likely be told, “ Yes, that is acceptable under the Regional Plan.” A user who wants to liquid-cool a nuclear reactor would probably be told, “No, that is not acceptable under the Regional Plan.” The council would not need to give a quantity nor a consent period. The council would only need to provide oversight that the users are following the law, and that the market manager is appropriately meeting environmental protection constraints such as aquifer drawdowns. The market manager would calculate the best allocation in the water auction, taking into account all the environmental effects. The Forever Fair system would charge a sufficiently high price to ensure protection of the environment. The council would have the authority to shut down a market manager who does not operate the market properly, such as failing to meet the environmental standards.
A new user would purchase all his water for a probationary period, perhaps a year, on a pure “user pays” basis. The money he pays would automatically go to existing users, compensating the existing users for the water they have given up to the new user. In the second period (probably the second year), the market manager would give the new user a consent equal to a proportion of water the new user actually used (following the “use it or lose it” principle). The new user can then use this amount of water for free, can sell from it, or can buy in addition to it. This is the buy-in process.
A user who decides to end his use of water would be motivated to sell the water over a year. The money gained effectively reimburses the user for his buy-in money. In the next year, the market manager sets the user’s consent to zero, because that was the amount of water used. This again follows the “use it or lose it” principle. This is the “sell down” process. We have noted some concerns that users may not be fully incentivized to sell down in all scenarios, but the Forever Fair system is sufficiently robust that adjustments to incentives are easily made.
Admission to the trading system is therefore voluntary. Any user can get a resource consent under the existing process, or can obtain a consent through the “buy up” process. Furthermore, any user with an existing consent can choose to join the trading system or can leave the trading system, whether they have a consent under the old system or under the new system. If users do not believe they will be better off with trading, they do not have to trade. The manager will tell each user the price of water in their area. If the user thinks the price is good, they can buy or sell at the next auction.
One implication of this system is that all users must have water meters. The water market manager may choose how to monitor water use and assess penalties in order to ensure that the water market manager meets environmental obligations.
Another implication is that the trading system does not need to be run by a governmental agency. Trading could be run by a private cooperative, with government oversight only for environmental standards. This would allow the quantities of water bought and sold to be kept confidential (except where tax issues arise). Only prices would need to be public, in order to give current users, prospective users, and the public correct signals for the value of water.
Further implications are discussed in the form of questions and answers in an Appendix to this report.
The current research has focused on allocating groundwater from a number of groundwater takes from an aquifer; however, further development is possible for the technology.
For example, the system has the potential to be expanded to include water quality as well as water quantity. Discharges can reduce the ability of society to meet environmental water quality objectives. They also decrease the value of water to some users. The problems that discharges cause to the environment and the community are examples of economic externalities, where third parties are injured by users’ behaviours. A market system that gives correct price signals to users about their discharges will give users economic incentives to reduce their discharges.
The requirements of a market system considering water quality are identical to the system for water quantity: a hydrology model, a web-based bidding system, and an operations research model that allocates consents and sets prices, while guaranteeing that environmental standards are met. Users would buy or sell from their consents.
Although the underlying mathematical programme would change, the system has the potential to expand and consider water trades in mixed surface/sub-surface water use systems.
Dual markets for both fresh water and for discharges would have enormous benefit for the economy and the environment.
The software development has led to the development of the core software required for implementation of a web-based trading system. The key new software is an Internet server—we used the Microsoft program Internet Information Services for XP. The server we developed will maintain a register of users, accept bids from approved bidders, and place their bids into a database. When the time for an auction arrives, the server calls MODMAN to develop a steady-state response matrix for the aquifer. MODMAN uses MODFLOW to identify the influence that a unit of water abstraction at each well has on all water users (and key environmental locations). The response matrix is used by MODMAN in a modified LINDO linear programming program along with the constraints to optimise the social benefit of water use. The decision variables are the amounts of water taken from each well. The output of the programme is the amount of water extraction allowed for each well and the price of water at each well (and at other places in the aquifer of interest). The Internet server then notifies each of the users of the results of the auction and allows bids for the next auction.
The user interface developed for users is shown below (values used are fictional).
The programme uses a unique identifier for each water user (“well05” in the example above). In the case above, the auction has taken place in time of extreme drought when all consents have been reduced to 17% of their non-drought values. This has reduced this user’s water allocation from 1ML to 0.17 ML. Each auction has its own identifier and the user can track previous bids and how much water they have traded and how much money they have paid to or received from the market manager.
To manage the auction information, we have developed a series of databases within Microsoft Access for the Internet server to use. Below is the database page for the auctions themselves. The system has a date and time when the auction occurs, the option of whether the auction is final or not, and also the option to impose an across the board limitation on water availability in times of drought. The database also stores the net income or cost (if any) to the auction manager.
A database of the users has also been developed. Users must submit a password and the correct ID to enter the system. The database retains this information to check against. It also can contain a number of other fields that could be useful for the market manager and the Internet server.
A great deal of code has been developed to make the system work; however, it is not easy to display or transfer. Because the software is its own Internet
server, the software cannot be readily added onto pre-existing networks. We found that a number of computer networks on the University campus would shut down
this application package because of its security threat. The software was trialled in network mode with about 10 users on
The University of Canterbury has used a grant from the NZHS to further develop a technology to allow trading of fresh water. This technology is called the Forever Fair Consent Trading System. It has the potential to improve the environment and the economy. Based on users’ bids, the hydrology, and environmental standards, the system allocates water to users and determines prices at every location. Users get accurate price signals to maximise the economic use of the water.
The grant has allowed the development of the core software needed to implement the technology. The software is linked to its own webserver and so is not easily downloaded. Please contact John Raffensperger regarding the potential to receive the most recent software package of the University of Canterbury effort ??. Development work remains to use realistic hydrogeologic conditions and improve the user-friendliness of the software.
The use of a trading system has the potential to simplify the existing contentious resource consent process. Users’ risk and uncertainty would be greatly reduced, because users could be surer of the costs and time required to obtain water. The technology shows potential to be extended to manage surface water in rivers and canals, and discharges such as nitrates.
Not much better than they can now. Obviously, with no water, there is no trading. With a little water, prices would be high. Our system is designed for general water shortage (certainly not for flood conditions), where some water is available, and where the environment must be protected. In these conditions, users can be made better off by trading, but the trades must be managed properly.
Under the Forever Fair system, the market manager is empowered to adjust resource consents, especially to lower them proportionally in time of drought. These proportional consents can be calculated by our system.
Trading is voluntary. If users wish, they can sell some of their allocation, and make money at the high bidder’s expense. Under the “user trades” configuration, all the money spent by the high bidder goes to existing users.
This is a problem of enforcement, not one of market design. We have not designed a penalty system; enforcement may best be handled by government. Traders must meter their water, and use no more water than they have been allocated after trades.
The Forever Fair system requires information on the flows that are needed to satisfy environmental standards. Government and community, with help from hydrologists, choose these flows. Examples include head difference constraints to maintain flows in certain directions, draw down constraints to protect the aquifer, minimum flows in surface water bodies, and several other types.
Will a user be penalised for being close to a sensitive region? Yes. Prices will be higher near sensitive places. That is good! We want to send price signals to users that they should reduce water use near environmentally sensitive areas.
What about the effect of water use over time? Water use now can hurt the environment in the future. The Forever Fair system simulates future water flow; a price may be high now to prevent future problems.
To a great extent, yes. In fact, we can allow trading between two owners who have individual consents for one well. In general, this depends on the time scales of the simulation and the auction. The Forever Fair system would do a better job of managing such trades if the auction were run relatively frequently, and if the simulation were run at a time scale that identifies the effects of the water flows.
The costs to set up the Forever Fair system includes the information technology for trading (for a secure web site, financial transactions, etc.), the hydrology model, and setting the environmental standards. On-going maintenance of the hydrology model requires operating costs. Water use requires monitoring and enforcement.
The research literature says that trading is much cheaper than augmentation by a very large factor. To get the most improvement per dollar, New Zealand should invest in trading before investing in augmentation, as trading should be both economical and highly effective. This contrasts with new races, canals, pumps, and pipes, which are effective but expensive.
If the hydrology modelling can be done, then we can develop the market. Note how this narrows the problem of “How do we allocate water?” to “How can we measure the water flows in this region?” We have simplified the problem from a complex one to something that is much simpler. Furthermore, we believe an approximate model would allow trading, and would improve the existing system. The government is already allocating water based on hydrology models. Adding bids to this can only help. Prices would tell the market manager which areas require the most accurate modelling.
Our system is for businesses that use large quantities. The focus is on large quantities of water used as an input to business, not about small quantities used by households.
A rich player would have to buy all the land, too, in order to use the water. This seems unlikely.
Possibly, but if a hoarding user sells when the price is high, that will help lower the price at that time. The result in the long run is that people will not hoard water, anymore than people hoard bacon.
It is true that trading water is not like buying lollies, and we agree that no one owns it. Water cannot be privatised, because anyone can dig a well and draw ground water. Whether or not water is privatised, decades of research have shown beyond doubt that fair and efficient allocation requires a system of prices. Water has not been traded before, because New Zealand had a surplus of it. Government could reasonably give a consent to anyone who could show good use. Now, about 75% of New Zealand’s water is used for the business of agriculture, and water supplies are becoming short, so better management is a necessity.
Trading water is more complicated than trading lollies, because a trade between two users will affect other users and the environment. We now have technology that accurately measures these effects, and can allocate water based on the hydrology, the price that users are willing to pay, and the environmental standards. Our market system allocates water in a way that protects the environment and maximises the public welfare. In short, the best for the community is a market system based on environmental standards.
Water trading in other countries has sometimes been an improperly regulated free-for-all. Problems have occurred in Chile and Australia. The environment has suffered, and some people have been injured by others trading.
However, ours is different, because we start with quantitative protections for the environment, such as desirable stream flows and prevention of coastal salt-water intrusion. Roughly speaking, our system first gives water to the environment. After the environment is protected, the remaining water is allocated to the highest bidder, as adjusted by the hydrology.
District councils will still make sure that water is used appropriately, and the community will continue to have a say. If the use of water changes to a higher economic use, this will make New Zealand more competitive. As long as the use follows the law, “undesirable” should be a use that is of low economic value. In other words, “undesirable” is the same as either “illegal” or “waste”. The former is a matter of legal enforcement; the latter is a matter of market forces.
Yes, this could happen with the current consent process. We believe in the “use it or lose it” principle. A user who sells water should have their consent adjusted to a proportion (75% or 100%) of what they actually use. In this way, a user who makes a windfall by selling water in one year will have their consent automatically reduced to what they actually used. The water they sold would be reallocated to all users based on their bids.
By adjusting resource consents to a proportion of actual use, people with large investments are reasonably protected. At the same time, sellers are compensated on a one-time basis. This is what we call the “sell down” process.
This is like saying you will waste bread if you can buy and sell it. If people can easily buy water, then they will not mind selling, because they can reasonably be assured of getting water when they need it. It would be a foolish person who would waste water when the person could get money for it instead.
With a “buy up/sell down” scheme, a new user would first get council approval for the type of use. In the first period (probably the first year), the new user would buy all his water on a pure “user pays” basis. The money he pays would go to existing users, to compensate them for the water they have given up.
In the second period (probably the second year), the market manager would give the new user a consent equal to a proportion of the amount of water the new user actually used (following the “use it or lose it” principle). The new user would be able to buy in addition to the new consent, or sell from it.
This would work if the hydrology were uniform across all of New Zealand. Unfortunately, some places are naturally wet, some places are naturally dry, and some places are environmentally sensitive. Furthermore, users’ needs change over time, so it is in their best interests if they can trade with each other.
In our research, we have found that even if all users bid identically, prices will differ from well to well. The reason is that the hydrology differs from well to well. We do not expect the price of fertiliser to be the same as price of lamb, because they are different commodities. Another analogy is transportation. Transporting wine from Kaiapoi to Christchurch costs less than transporting wine from Auckland to Christchurch. Thus, the value of water will be different in different locations.
New Zealand would be much better off if the price of land, the price of water, and the price of discharges were all separated.
We think that consents should be set to a proportion of the previous year’s actual use, following the “use it or lose it” principle. Therefore, if you are using the water, your consent will continue from year to year. You have no need to renew after X years, because you can simply continue your existing consent permanently, as long as you do not voluntarily “sell down”, as long as you have an acceptable use for the water, and as long as you follow the law.
The consent process would be greatly simplified with the “buy up/sell down” scheme. You first get council approval for the type of use, and then simply buy the water you need. In the next year, your consent becomes a proportion of what you actually used.
In drought, council already can turn off your supply with little notice, and you have no recourse no matter how much you are willing to pay. Our trading system would be far more flexible. It would let some water go to those who really need it, as demonstrated by their bids. In addition, we have developed the scientific capability to adjust consents proportionally, based on the hydrology and available water.
It depends on why your consent restricted your irrigation. If it is only a quantity issue, trading will most likely allow you to get more water, if you are willing to pay for it.
You are not required to trade. If you do not trade, then after each auction, you can see the price for water in your area. If you think the price is very high, you could sell water at the next auction and make money. If you think the price is very low, then you can buy water at the next auction and make more money from your farm. The flexibility to buy or sell will reduce your risk.
Under the “buy up/sell down” scheme, new traders would compensate existing users for the water they take away. The only way new users could come into a fully allocated market is by using water more economically than existing users. If that happens, it means that the water is going to a better use, which will generate more benefits for New Zealand as a whole.