AG WATER PRICES: Groundwater Externalities and the Agricultural Response to Water Pricing

Dr. Ellen Bruno is a Cooperative Extension Specialist in the Department of Ag and Resource Economics at UC Berkeley.  At a recent Silver Solutions webinar, she shared some of the preliminary results on a paper she is working on with colleagues Katrina Jessoe at UC Davis’ Ag and Resource Economics Department,and Michael Hanemann with Arizona State University who is also a professor emeritus in the Department of Ag and Resource Economics at UC Berkeley.  The study considers the impacts of agricultural water pricing and the effect on water use and land use change. 

Common pool resources and the tragedy of the commons

The motivation to do the research was the concept of the Tragedy of the Commons, which was popularized by Garrett Hardin in 1968.   The tragedy of the commons is a situation in a shared-resource system where individual users, acting independently according to their own self-interest, behave contrary to the common good of all users by depleting or spoiling the shared resource through their collective action.  Or simply put, anyone can access the resource, and one person taking advantage of the resource means that there’s less available for other people.  These are commonly referred to as common pool resources.  An example would be fisheries; everyone can go out and catch fish, but if one person catches a fish, that means there’s no longer that fish for other people to catch. 

Generally we think that the incentives for an individual using the type of resource are not well aligned with what’s in the general public interest and what would sort of maximize the well being to everyone in society that benefits from that resource,” said Dr. Bruno.  “In the absence of either regulation, privatization of the resource, or some sort of Elanor Ostrom-style local community governance, that in most cases when you have a common pool resource, it can lead to overuse of the resource and in some cases, the collapse of the resource.”

Groundwater is a common pool resource, and in many places, the property rights for the groundwater itself are not well defined. We don’t think of this as a private resource, but at the same time, it’s not regulated in many cases and oftentimes not even metered, and this can lead to issues of overextraction of the groundwater resource, Dr. Bruno said.

So if Farmer A pumps a lot of groundwater, he/she might notice that the cost of pumping is increasing as the water table lowers, raising the energy costs of extraction, but it might also increase the cost to Farmer B next door, but Farmer A is not necessarily taking that into account.  This is an example of a negative externality, which is a cost that is suffered by a third party as a consequence of an economic transaction.

We think this has contributed to the reason why many of our groundwater aquifers worldwide are under stress,” Ms. Bruno said.  “There are externalities in several dimensions when it comes to groundwater.  It’s not just making the energy costs greater for my neighbor, for example, but it might mean that it makes it harder for any individual to plan for the availability of the resource in the future if anyone can come in and extract that resource.  You can’t assume that if you don’t extract it today that it could be available in the future.  There are also water quality externalities too; it’s not just about the quantity of the groundwater but one person’s pumping could affect the quality of their neighbor’s groundwater.”

All of these groundwater externality issues could potentially be exacerbated by climate change.  With increased temperatures, both the form and the timing of precipitation are changing, and in California, this means less snow, more rain, and the snowpack melting earlier in the season which will complicate how runoff gets captured in reservoirs and used later in the year.  Climate change is also expected to create conditions of more variable precipitation with more frequent and severe extreme events like droughts and floods. 

Groundwater is a really important buffer to variation in surface water supplies and rainfall, so when we have more frequent and severe droughts, there’s going to be an increased reliance on groundwater supplies,” said Ms. Bruno.  “We’ve seen this historically in past droughts that farmers have relied on their groundwater supplies to make it through those changing conditions and so it’s all the more reason to be thinking about why we should manage groundwater and think about the use and allocation of this resource.”

Economists think that by taxing the resource, it will correct some of these negative externalities of common pool resources.  The idea is that if there is a charge to use the resource, people might change their behavior such that it’s more aligned with what would be in the public interest and correct these instances where the individual incentive is not well aligned with what’s good for overall society. 

Agricultural groundwater is typically not priced; there are energy costs related to pumping the groundwater up from below so it’s not completely free, but the actual groundwater itself is not priced, she noted.  So a policy maker or an economist might think of pricing as a policy instrument to correct these consumption externalities.

Groundwater pricing might also be important in the context of technological innovation.  “Prices spur technological change, and therefore the absence of a price would dampen the incentive for technological change,” Ms. Bruno said.  “We’ve seen empirical evidence of this happening in other sectors, like when we see an increase in energy prices, there’s an increase in innovation related to energy efficiency.  It’s the same with carbon.  People have studied carbon markets in the EU and how that’s driven low carbon innovation.  So pricing of water can also be important when it comes to thinking the adaptation to climate change and potentially increasing water scarcity in the future.”

So we’re thinking about water pricing as potentially serving as both a corrective policy instrument for the overconsumption associated with groundwater being a common pool resource, and also as an adaptation strategy to climate change and changing water supplies.”

Challenges to the study

The study is looking at a specific irrigation district that happens to already charge a water pumping fee for agricultural users to try to examine empirically what the presence of a water price has done for water use and land use change of those agricultural users in that particular region.  The study will estimate the effect of the volumetric groundwater pricing on a suite of different behavioral changes, such as agricultural water use, land fallowing, crop switching and the conversion of land out of agriculture entirely.  The goal is going to identify the causal impacts of these water prices on the different metrics.

She pointed out that this presents a challenge in a non-experimental setting.  Ultimately they want to understand how water pricing actually causes these different factors to potentially change to understand how farmers respond to the introduction of groundwater pricing, but it can be challenging to get data on wells for agricultural groundwater extraction as they are often unmetered.  It’s also fairly uncommon to see volumetric water pricing for agricultural groundwater.

It’s also difficult to understand what the appropriate observational unit is in this study.  They are trying to study grower-farmer behavior and so ideally they would want to know which parcels of land correspond with each other, how much water use is occurring at a particular well, and what parcels of land is that well watering, but it can be difficult to map water use to particular land parcels.  Finally, they are using observational data to try to come up with a causal estimate, so working within a non-experimental setting but trying to identify causality, if possible.

However, in this particular empirical setting, they were able to overcome some of those challenges to get closer to understanding how water prices affect water use and land use.  The study area is the Pajaro Valley along the Central Coast, which is a very productive agricultural region of about 30,000 irrigated acres known for strawberry production, as well as other berries, apples, grapes, artichokes, lettuces, and other vegetable row crops.  Importantly for this study, almost 100% of their water supply is from groundwater so the impacts can be isolated.  The water agency collects a lot of data and charges a volumetric price for agricultural pumping.  And it’s a ‘quasi-experimental setting’ with a treatment and control group, although non-random, but it’s a setting that will help move away from correlation and try to move towards causation.

Ms. Bruno noted that the reason this is important is that agriculture as a whole is likely to be hit pretty hard by climate change and changing water supplies, and as a sector, they are responsible for a large fraction of the global water withdrawals.  There are a lot of irrigated crops worldwide that are important for the food supply.  Groundwater itself is an important component of the water supply as it is an important way to buffer other variability in rainfall or surface water. 

In California, the Sustainable Groundwater Management Act (or SGMA) provides a statewide framework for groundwater management over the next several decades to regulate some of the historical groundwater overpumping throughout the state of California and the associated negative impacts, such as declining water tables, land subsidence, and deteriorating water quality.  SGMA requires all medium and high priority groundwater basins to develop plans to manage their groundwater basins sustainably.  These basins account for the vast majority of the groundwater pumping in the state.

The local management agencies in charge of these plans have been given a lot of flexibility and authority in how to achieve groundwater sustainability, including the ability to leverage groundwater pumping fees, so this research directly has implications for SGMA and pumping fees that might be implemented throughout the state, either to raise funds to do supply augmentation projects or to do demand management itself,” said Ms. Bruno.  “In either case, understanding how water prices are going to change the agricultural landscape is going to be important in the context of SGMA.”

The context of this study

She presented a map of all the wells in the dataset with each black dot being a groundwater production well in the Pajaro Valley Irrigation District; the service area is outlined in blue.  There is quarterly extraction data for each of the wells.  The time period for the study is 2005 through 2017.

For the ideal experiment, if we wanted to look at how pricing affects water use, we would want to randomly assign a subset of these wells to a higher water price and then compare how their water use would differ from those who didn’t receive the price treatment,” said Ms. Bruno.  “But this is a real life context where we can’t randomly give some subset of people a higher price than others, and so we have to work with what we have.  In this case, we’ve got the next best thing, where all the wells that are in this purple region that are located near the coast and halfway through our sample period, they received a higher price.  So it’s not random as to who received the higher price, but we do have a treatment and a control group, and we have a pretreatment and a post treatment period, and so we can exploit that, especially in combination with institutional knowledge about what created this separate zone and determined who got treatment to try to understand the impacts of prices on water use and land use.

She next presented a graph of the price treatment, noting that the wells that are in the purple region are in the “delivered water zone” or DWZ.  Prior to the third quarter of 2011, everybody was charged the same price to pump water as a price per acre-foot.  Then in the third quarter of 2011, those inside the delivered water zone received a disproportionately rate increase that continued afterwards.

We’re going to exploit that jump in prices that was only to those inside the delivered water zone to understand how the price increase affected water use and land use,” said Dr. Bruno.  “We’re coupling this with annual land use data that the water district has conducted.  This was created through field surveys and we have it for a subset of the years that we have pumping data for.  It tells us the crop composition and delineates what’s irrigated versus non-irrigated agricultural land, what’s fallowed, and also what is not agricultural lands so we could distinguish residential or open green space from ag land.”

Dr. Bruno also noted that the delivered water zone is right next to the ocean so there are certain properties about those farms that might be distinctly different from the farms that are in the rest of the valley.  There is a monitoring network of about 200 wells throughout the valley, shown in purple, that measure different water quality metrics, one of which is salinity, the most important in this context due to the proximity to the ocean and the historical overdraft in this area.  There is also information on the depth to the water table, which is important things to control for in the setting, since those who received treatment were not random.

The last challenge was matching a well to what crops were being watered, which is important for understanding what the observational unit is, so they used the county tax assessor property boundary data, and overlaid those property boundaries on to the land use and aggregate water use within those parcels and treated those property boundaries as a farm unit in the analysis going forward.

Our plan is to compare the water use and land use separately across these two pricing zones before and after the price change,” said Ms. Bruno.  “It’s the comparison between the treatment and the control group.  The treatment group are the pumpers that received a price increase in the third quarter of 2011, but because this is a non-experimental setting, it’s really important to understand what determined the treatment, how did certain farms get assigned a higher price and those others did not, what is the delivered water zone, what makes that zone special and what’s driving this price change.”

She then gave some background on why the group received the higher price as this informs the statistical regression.  “We want to control for these things because ultimately we want to argue that conditional on the differences, that assignment is random.  Essentially we need to make an assumption about the counterfactual world of those pumpers inside the delivered water zone, if they didn’t receive the price increase, what would have happened to their pumping that we cannot observe.”

The reason behind the delivered water zone was that the water agency wanted to develop recycled water supplies and so they initiated the pumping fee in order to raise revenues to fund the recycled water program, she explained. The recycled water program was created to help reduce some of the seawater intrusion problems that were happening due to the large amount of groundwater pumping occurring in a coastal basin.   The recycled water itself is a scarce resource, only about 3% of the basin’s total water use, so they decided to allocate it to only some of the pumpers that were close to the ocean, because there were great benefits to the basin overall if those pumpers who were close to the coast substituted some away from their groundwater pumping and use the recycled water.

We know that those inside the treatment area are receiving recycled water, and that’s a very important control variable in our regression that we have data on and we can control for directly,” Ms. Bruno said.  “But there’s also reason to believe that those wells might differ in terms of the groundwater quality coming out of those wells, will differ in their water quality as well as potentially the depth to the water table that they’re drawing from.”

With respect to the timing and what drove the price split, Ms. Bruno said her understanding was that it had to do with a Prop 218 lawsuit where someone outside the delivered water zone sued, saying it’s not fair that all were paying the same price even though those inside the delivered water zone are the only ones receiving recycled water deliveries. 

As a result of that lawsuit, they decided to make differential prices where those inside the delivered water zone were going to pay more because they are receiving recycled water, but the actual timing of when the treatment came into play has to do with the timing of settlement of this lawsuit which went on for a bit, and so we can sort of essential safely assume that the exact timing of when the treatment came in was random.”

The slide at the lower left shows the raw quarterly groundwater extraction data from well level data; the dotted line notes the third quarter of 2011 which was when the price treatment began.  Dr. Bruno noted that while you can’t tell a lot from the raw data, it does seem like something different is happening after the price change.

The study is doing a regression analysis for both water use and land use in three separate equations.  The y variable on the left hand side in the equation on the upper right can be several different things, but in this case, it will be water use, irrigated usage, and total agricultural land.  The parameter of interest is the Delta parameter which shows the differential effect of those inside the delivered water zone.  Other parameters on the right control for parcel level fixed effects, aggregate time varying shocks which is anything that happens to everyone in the basin at any given point in time, and the control variables such as the recycled water deliveries, groundwater quality, depth to the water table, rental rates for the farmland, and things that could potentially change at the region level over time that might be correlated with treatment and are outcome variables and if excluded, could potentially bias their estimate of the treatment effect.

Preliminary results

Dr. Bruno than previewed some of the preliminary results. 

Starting with the impacts of prices on water use, they found that the price increase, which was about a 30% increase, led to a 28% reduction in groundwater extraction.  “This result was statistically significant and robust across several specifications, robust to the inclusion and exclusion of these different control variables as well as some other tests,” she said.  “We think this is a pretty large impact; I was surprised at the magnitude of the effect that we’re seeing here, so a fairly large economically meaningful impact of water prices on water use.”

There were also fairly large decreases in total agricultural land and irrigated acreage as a result of the price increase.  “We’re seeing that the 30% increase in prices led to a 19% decrease in total agricultural land and a 14% decrease in irrigated acreage,” she said.  “The impact on the total land in agriculture so that’s including fallowed land is greater than the impact on irrigated acreage and so within agriculture, the fraction of land that is irrigated actually went up as a result of the price increase, and that’s what we’re seeing in the data.  Therefore, if you were to look at fallowed land just as a percentage of land in agriculture, it would be declining as a result of the price increase.”

This is something that was maybe a little counterintuitive when I first saw these results, but then in thinking about it more, I think it can be justified by thinking that in the face of a permanent price increase, if you were going to make some land adjustments, you might be more likely to sell off or permanently convert your least productive agricultural lands and maybe irrigate a greater portion of the land that remains in agriculture, if the price is perceived as a permanent shock.  I think it’s an interesting result.

She reminded that these are all preliminary results.  They are also looking at how the price increase affected planting and different crop categories.  The table is by crop category. 

What we’re seeing so far is that agricultural water pricing is having significant impacts on both agricultural water and land use,” said Dr. Bruno.  “We’re seeing large reductions in water use, we’re seeing decreases in irrigated acreage, decreases in the total amount of land in agriculture, and some crop switching as a result of the price increase.”

The results will be important in the context of SGMA as agencies across the state try to figure out how to correct overdraft.  The chart is from a recent PPIC review of all of the groundwater sustainability plans in the San Joaquin Valley, where they found that agencies on average are saying about 80% of that overdraft is going to be met with supply expansion projects, and the remaining 20% through demand management, such as restrictions on pumping where they might consider a price or tax on pumping as an instrument for demand management.

They conclude that this is pretty optimistic to think that 80% could come from supply expansion,” said Dr. Bruno.  “Demand management is an important part of meeting the requirements of SGMA, in which case understanding the role of prices could be directly informing impacts in this demand management space, but even for the supply expansion as agencies are talking about implementing fees in order to raise revenue to fund these supply expansion projects, so even if the pumping fee is not put in place for the purpose of demand management, it still can have impacts on water conservation and land use change that I think will be important going forward in the implementation of SGMA.”


QUESTION: Have you considered that the total decline in agricultural land could be due to other factors apart from the price, such as increasing urbanization, that would be different in the coastal area as well, where land prices are higher?

ANSWER:In terms of residential pressure, right now we only have the farmland rental rates included in there, which is a regional variable, so to the extent that that varies and relates with, the farm land rental rates will be correlated with residential land value as well, but that is something we have talked about and something we need to try as well is getting some measure of that pressure on residential demand more directly to include that variable.”

QUESTION: Given that SGMA is developing right now, how do you see this other water districts or irrigation districts to use economic tools like pricing to achieve reductions in groundwater use?  Do you see if they are going to apply it?  How popular are economic tools like pricing or taxes?

ANSWER:My general understanding is that pricing as an instrument itself is not a very popular one, although several agencies that I’ve read parts of the plans of are talking about leveraging fees in order to raise revenue, so it’s for a different purpose but might still be meaningful in that context as well.  I do think economic instruments in general have received some attention in the context of SGMA in that several groundwater management plans talk about the potential for groundwater trading and markets for groundwater, and so it’s sort of a broader answer to your question, I think there are interests in economic instruments, less so on price directly or taxes directly as a corrective instrument.

QUESTION: Early in your talk, you alluded that increased pricing might bring more innovation, that might help save water and I’m just curious if you noticed any of that where the farmers are trying to be more innovative to maximize their water use instead of letting the land to go fallow.

ANSWER:It’s something that I’ve been very interested in but I find it challenging to get information on that at the farm level, how these farmers are making adjustments in terms of irrigation efficiency.  It’s something that I’ve looked for in this area and we’ve gotten some county-wide statistics on irrigation efficiency, so we might be able to make some comments on sort of trends at the county level, but we won’t be able to pin down the same sort of empirical estimates that we’re doing on these other factors, because we just don’t have the same data on irrigation efficiency at the farm level.”

QUESTION: What would be your dream data?  If you had the data to assess irrigation efficiency, what would you want that to look like?

ANSWER:  “Ideally, knowing what sort of irrigation technology that is being used at each of these farms, at each individual farm, because that’s what we’re working with here and the reason we like having the most micro-level data as possible is because then we can throw in a bunch of control variables that can soak up things about the individual farms that are unobservable to us, like there might just be differences in the ways farmers farm that explain differences in outcomes that we can’t possible observe, and so it’s helpful to have the most granular level possible.”

QUESTION: When will the report or paper be available?

ANSWER:I will have a working paper available very soon.  In economics, it’s very common to share working papers because there are huge publication lags and anyways, it will be years before it’s in actual publication, but a few months and then I’ll have something I can share with you all.”

Print Friendly, PDF & Email