Applying AI to the California Water Crisis?

Article continued from previous page.

Regulation by futures exchanges

To be enforceable, these contracts do need to be placed on exchanges (“self-regulatory organizations” or SROs in finance speak.) Open interest on exchanges should be made public and limited to prevent excessive speculation (a source of a potential systematic shock), and contract terms standardized by the SRO exchange to facilitate an efficient market. Futures are a form of derivatives. Requiring use of an SRO for non-insurance contracts and publishing & limiting open interest eliminates most of the problems associated with derivatives contracts. It allows enforcement of rules generally designed to prevent excessive open interest, “cornering” of the market, extreme price swings, and other irregularities. These regulatory systems have been well-understood in the United States for well-over a century (although appear to have been somewhat forgotten when Enron was trading in electricity futures).

So this is the “dumb” aspect of futures markets. As shown, this would already help farmers more efficiently select water-efficient crops. It would do this by aiding price discovery, allowing farmers to efficiently coordinate with one another in their selection of crops and water purchases.

Futures markets would also help on the water supply side. We listed a large number of water technologies. Having a waters futures will help select the best technology at any given time.

The invisible hand of futures exchanges

For example, conversion of ocean-going oil tankers to instead transport water from Canada is a fairly elastic technology. These oil tankers can quickly be brought on-line, as has been done in the past. A typical well-regulated water futures market can easily provide liquid water contracts for a year or two. (US financial markets, as well as prediction markets, have experience with longer-term price discovery products such as LEAPS or multi-year predictions contracts on New Zealand’s iPredict markets. Unfortunately, these products, while generally commercially viable, then to be less liquid than near-term contracts.)

Thus, as the future price of water begins to rise, an oil-tanker company could use the Futures market to lock in a transportation contract. (Generally, this would involve constantly comparing the future profits on the existing CME GLOBEX oil deliveries contracts with the future profits from water delivery contracts & factoring in the small conversion cost to switch between oil to water.) An oil tanker company could “lock-in” a profit for many water runs. Once committed to the water delivers, the future price of water would fall, enabling uncommitted farmers to switch to more water-intensive (and more profitable) crops. In this way, a waters futures market facilitates coordination between ocean-going oil/water tankers and farmers.

Given the caveat about longer-term liquidity of futures, the futures market would also facilitate investment in longer-term projects. Critical decisions need to be made whether to invest in desalinations projects or in building aqueducts from places like Idaho. By providing investors with a way to “lock-in” long-term profits from water deliveries, water Futures and water price-prediction markets provide a means of financing these projects. (Investors considering building an aqueduct short the future price of water for several years of deliveries, which should already provide most of the financing for many smaller projects such as desalination plants.)

Energy and desalination as “long-term” exponential technologies

We’ve already pointed out that, under the Kardashev Civilization scale (presumably well known to the folks at Singularity U sponsoring the original contest), energy and civilizational progress are closely linked. This is also largely true when considering water technologies. Desalination, in particular, is energy dependent. Thus, in addition to crop and water markets, California should also invest to ensure that our historically controversial energy futures markets remain liquid, functional, and competently regulated.

A small desalination investor would be able to finance a desalination plant by going long electricity futures (the essential raw material for desalination) and shorting water futures. Thus California should have water, electricity, and crop futures markets. With well-functioning markets in those areas, farmer should automatically allocate water-efficient crops, and at least small desalination plants should build themselves (in the absence of more economical water-delivery technologies).

The role for government is not to ration or centrally plan, but to facilitate that these markets operate smoothly, and police abuses such as un-permitted water withdrawals from politically privileged but unethical players. Some planning may be required for long-term projects where Futures markets are too illiquid for efficient price discovery. Even here, however, an illiquid futures market can provide some guidance; planning and financing can still be facilitated by private entities, albeit a bond issues might be needed in place of shorting futures contracts.

Futures markets and Futarchy short-term exponential technologies

Finally, we can begin to explain how futures markets are “exponential.” We’ve already discussed Futarchy. As artificial intelligence becomes more sophisticated and reliably takes over more trading, liquidity (through arbitrage by AI systems) will become more sophisticated. (For example, trading systems have long applied artificial intelligence techniques natural language processing to news releases and twitter markets to automatically place trades. Weather futures traders presumably use sophisticated weather models, while some oil traders are known to use sophisticated oil inventory modeling. The price discovery from these models would feed into water and crop futures markets.) As a result, water, electricity, and crop contracts become liquid further into the future (together with an appropriate regulatory infrastructure that ensures these contracts are honored). As this happens the system becomes exponentially more intelligent. It can make longer-term decisions. It can invest efficiently in more expensive but longer-term technologies, such as multi-year aqueduct projects.

One key to Futarchy is public confidence that the futures prices being set many years in advance are accurate. In general, these prices will be set by sophisticated, and increasingly complex, “black box” computer algorithms at proprietary trading firms. As we have argued elsewhere, to insure confidence, the government will need to invest in public computer models (through e.g. university research). These public models provide a baseline against which the general public can check the more sophisticated proprietary models for reasonableness. They also form a starting point on which private corporations can build these more sophisticated proprietary models. The price difference between the academic model and hopefully more accurate proprietary “black-box” models creates the profit incentive to invest in sophisticated exponential technologies like Artificial Intelligence.

Additional types of futures

Ultimately, one can imagine weather futures. (These already exist: http://www.cmegroup.com/trading/weather/. What is being proposed are more detailed weather futures for California.) Sophisticated computer models would influence these weather futures. Long-term weather futures (at a very granular level, such as aggregate rainfall) would also be possible, and influenced by modeling of El Nino, solar cycles, greenhouse gases, and other predictable long-term weather drivers. In this way, through the use of artificial intelligence, factors such as green houses gases would impact multi-year water delivery prices.

Via the futures market, farmers would be using sophisticated artificial intelligence models to decide on crop commitments, without knowing anything about the computer models. Moreover, weather modification and geo-engineering programs, once legal, political, and environment concerns are overcome, could be financed through these futures and Futarchy markets. (A weather modification investor would simply short an inverse-rainfall contract. The resulting prediction of increased rainfall, anticipating the application of the weather modification technology, would then also affect water delivery prices, and hence farmers’ decision making.)

Example California water future arbitrage trading strategies

1. A farm converts (“arbitrages” in trading parlance) water into crops. If the price of water falls below the price of a crop’s futures less the farmers’ cost, the farmer locks in a profit by planting that crop. Crop futures and the need for long-term agricultural planning and rail transport is what led to Futures being invented in Chicago in the 19th century. Farmers are thus very familiar with the concept of futures markets. As noted earlier, water futures are specific to regions of the country where water is expected to remain expensive and volatile in price; most regions of the country (and certainly not the 19th Century midwest) do not need water futures just yet. In modern farming, energy is another important input, so the farmer can also factor in energy futures (gasoline or oil) into his decision model.

2. A desalination plan converts (“arbitrages”) energy (oil or electricity futures) into water. As already discussed, for thermodynamic reasons, “energy” (entropy) will also be a significant cost in desalination. When the future price of water exceeds the futures price of energy plus other operating costs, the operator can lock in a future profit by operating the plant. For small enough desalination plants, a high enough future price of water may justify building the plan. “Small enough” may become increasing larger as the time-horizon for the futures markets moves further forward in time with increasing liquidity.

3. Similarly, diverting an oil tanker for water transport converts (“arbitrages”) oil futures into water futures. If price of water futures rises above oil futures, it becomes cheaper for an ocean-going tanker to deliver water rather than oil (from a source where freshwater is cheap). The oil tanker or barge operator can then lock in a profit by delivering water. (This arbitrage opportunity would also apply to highway tanker truck operators delivering water instead of oil or other liquids.) In this way, water transport strategies will compete somewhat with desalination facilities. Regulations surrounding export of freshwater as well as the ability of desalination to use forms of energy other than oil (electricity from renewable sources) may give desalination a longer-term edge.

4. Water and weather futures can be “stat arbitraged” or “pair-traded” against each other. Since at least some freshwater comes from rainwater, and rainwater futures already exist, sophisticated traders can potentially model the relationship and engage in what is called “statistical arbitrage.” In this way, the long-term weather models used to price weather futures will indirectly help price water futures. Incentives will be created for more accurate long-term statistical forecasts of precipitation.

5. Water can be stored (in reservoirs or tanks), and thus water futures expiring in one month can be arbitraged against another month. Water prices in peak demand months (e.g., summer irrigation season) will automatically be lowered through participation of reservoir systems in the futures market.

Conclusion

The establishment of liquid, multi-year weather, electricity, crop, and water-delivery futures contracts for different regions within California has some ways to go. The underlying technology, however, is very old. It is a tried and proven technology. Aspects of these futures markets for weather, electricity, and crops already exist. With addition of simple water delivery contracts, financing of small, short-term projects, such as ocean tanker delivery of water and small desalination plants, can already begin today. Moreover, these futures markets are already very heavily computer and artificial-intelligence driven. They are exponential technologies. As computers and AI improves, these markets will become more liquid. They will be able to make more accurate predictions further into the future. This will spur investments into the many other water technologies outlined above. Many of these water technologies are proven “old” technologies. Drip-irrigation, for example, has been around since the 1960s, but has not been used in California as has been historically been seen as too expensive. A futures market would encourage residential conservation of water, since municipalities would also interact with these futures markets, and offer rebates and market-incentives to residents. An efficient waters future market would facilitate long-term price discovery of water, enabling residential consumers and urban planners to make informed decisions regarding proven but expensive technologies such as urban dual-water supplies. With an exponential water futures and Futarchy system on-line, artificial intelligence would begin to guide much more efficient “exponential” multi-year planning and investments that would incorporate these many proven but largely unused “non-exponential” water technologies into an exponentially more efficient system.

Legal disclaimer: This is a journalistic opinion piece on the California water crisis. It is not intended as advice for your specific situation, or as financial or traditional engineering advice. Consult an appropriately credentialed or licensed professional before trading on the futures markets, constructing an aqueduct, or diverting an oil barge. See our terms of service for complete disclaimers.

1 2 3