American agriculture has long been driven by an awkward mix of free markets and regulation – with the latter a product of its own awkward mix of interest group politics, consumer protection and environmental concerns. California agriculture is hardly an exception.

Indeed, the state's giant farm sector offers a fascinating picture of markets and regulation at work, as California is buffeted by uncertainties about the cost and supply of desert-farming's two critical inputs: water and low-skilled labor. Arguably, the most surprising aspect of this never-ending tale is how resilient the industry has been in the face of rapid changes in politics, economics and the natural environment. Here's an update.

Water Roulette

First things first. California is home to just 2 percent of the cropland in the United States, but accounts for an eighth of total farm sales because it disproportionately produces high-value commodities: fruits, nuts, vegetables and horticultural specialties such as nursery products. All told, California's production in 2014 included $30 billion worth of fruits and vegetables – almost as much as the $31 billion of total farm sales of the second-largest farm state, Iowa.

Even though a drought began in 2012, California farm sales have risen each year. In 2011, California's last "normal" water year, farm sales were $43 billion. Sales rose to $47 billion in 2012 and $54 billion in 2014, the worst year of the drought. How was this possible? About half a million acres were fallowed in 2014 for lack of water and cheap labor, but this was land that would have been planted in low-value field crops.

In some cases, farm revenues go up (or at least not far down) during droughts because smaller output results in higher prices. However, there were only modest changes in the state's major fruit crops during recent droughts: the production of oranges fell 9 percent in 2014, while the strawberry crop grew by 2 percent. Likewise, the production of most major vegetable crops in 2014 was similar to the 2011-13 average, while tomato acreage actually rose by 15 percent.

It can be argued that California agriculture's greatest strength, the state's vast irrigation infrastructure, makes production less dependent on rain in the short run. But this system is also a singular weakness, since the supply of water to feed the network seems to be increasingly volatile. California uses an average 33 million acre-feet of irrigation water a year. (An acre-foot – the amount of water it takes to cover an acre one foot deep – is 325,851 gallons, the consumption of a typical suburban household and lawn for a year.) In normal years, 60 percent of it comes from surface water and 40 percent from wells. In dry years, these ratios are reversed, as water pumped from underground aquifers replaces water not available from dams and canals.

In 2014, California became the last Western state to regulate groundwater pumping, enacting laws that create local groundwater sustainability agencies to register private wells, monitor the water-measuring devices that must be attached to pumps, and regulate pumping to avoid exhausting supply.

The drought ended (or at least paused) in Northern California last winter; precipitation was normal. Indeed, at the beginning of April, the state's 154 major reservoirs held almost 22 million acre-feet of water, 85 percent of normal, so that the various entities with contractual claims to water received half or more of their regular allotments.

The California water system is powered by snow accumulation in the mountains of Northern California, with snow melting into runoff from the slopes and moved from north to south via the Sacramento-San Joaquin River Delta. Three factors shape the longer term outlook for water. First, most climate-change models predict warmer winters that are less suited to California's snowpack-based water storage. If more precipitation falls as rain rather than snow during the winter months, the mountains store less water. And without a way to store the runoff, the production of water-intensive but low-value crops, such as alfalfa (a kind of grass) for dairy cows, is likely to shift out of state.

Second, the elasticity of demand for irrigation water is slowly diminishing. This is because an increasing portion of the water is used to sustain long-lived fruit and nut trees, which can't survive a long pause in watering. Acreage planted in almonds, which normally need three to four acre-feet of water a year per acre, more than doubled since the mid-1990s, to 1.1 million. Meanwhile, land in cotton, which can be left fallow with only the loss of a single year's crop, declined 160,000 acres in 2015 – just one-tenth the acreage of 1980.

A third factor – one generally applauded by economists – is that more water can change hands in the private market. Water has traditionally been priced according to the cost of the government-built infrastructure that stores and delivers it – not, as competitive markets would have it, by the value of the water to the marginal user. Hence, most water for agriculture still costs relatively little. But with government-built contractual sources becoming less reliable and wells less productive as the water table falls, the (still thin) free market for water is recording sales that show some farmers are willing to pay $1,000 or more per acre-foot to keep trees alive. That, by the way, is three times as much as the big urban water agencies usually pay to obtain water for residential households.

Gov. Jerry Brown is proposing a tunnel complex to move water from Northern California around the Sacramento-San Joaquin River Delta and into reservoirs and groundwater aquifers in the San Joaquin Valley, the site of over half of the state's agriculture. If the tunnels are built (at an estimated cost of $16 billion) and water users are given more discretion to sell their legal allotments to the highest bidders, farmers who grow rice (needing over five feet of water per acre) and other water-intensive (but low-value) crops in the Sacramento Valley may well choose to fallow their land and sell their water to farmers who grow more valuable crops in central California.

One big question, then, is whether the state should invest heavily in water infrastructure to minimize the adjustments required of agriculture in an era of unreliable supply, or whether it should encourage water sales that shift water to its highest-value use with the existing infrastructure.

Government infrastructure development and greater dependence on markets define the ends of the spectrum of policy options. The neither-fish-nor-fowl status quo creates uncertainty for farmers and urban water districts, some of which are acquiring water for more than the price likely to prevail in a free market for the foreseeable future. So simply making the options clearer promises to make the water system more efficient.

Labor Supply in an Era of Nativism

Labor use in California agriculture rose 12 percent over the past decade, reaching some 415,000 (about 2 percent of the Golden State's labor force) in 2014. This increase reflects expanded production of fruits and vegetables that require harvesting by hand. The tilt toward labor-intensive crops more than offset the impact of mechanization in a few other crops, notably raisins. Labor's share of production costs ranges from less than 5 percent in field crops to 40 percent in berries, with an average 30 percent in hand-harvested commodities.

Some 72 percent of primary farm workers had only one farm employer, suggesting far more stable employer-employee relationships than the follow-the-crop migrant stereotype portrays. Fewer than 10,000 farm workers had five or more farm jobs in 2014, although some of the workers employed in processing and logistics may have worked on multiple farms. Farm employers face several challenges, including a statewide minimum wage of $15 that takes effect in 2022. That's likely to pinch employers far more in rural California than in cities. The $15 minimum wage is projected to be less than half the median wage in San Francisco in 2022, but 70 percent or more in the San Joaquin Valley, where half the state's farm laborers work.

What will happen to farm jobs when the minimum wage is 70 percent of the median wage? The optimistic scenario is that the San Joaquin Valley's fruit bowl will employ fewer, but higher-wage and more-skilled, farm workers. The pessimistic scenario is that the San Joaquin Valley will follow the pattern of Appalachia, remaining home to low-productivity workers who lack the savings, information and skills to move elsewhere.

The major farm labor challenge today is finding workers to replace those who depart. Almost 90 percent of California crop workers were born in Mexico and over 60 percent are not authorized to work in the United States. Most are men (75 percent) who have settled here with their families, and most are poorly educated. These workers are also aging: they have a median age of 39, very close to the median of 42 for all U.S. workers.

The share of farm workers who were in the United States less than a year was 5 percent in the early 1990s; that figure rose to 20 percent in 2000 when Mexico-U.S. migration peaked, and has since collapsed to less than 2 percent. Farmers are pursuing a variety of strategies to cope with labor supply uncertainty while waiting to see what happens next on immigration policy. These run the gamut from improving working conditions, to increasing productivity while reducing back-breaking effort with mechanical aids, to supplementing their work forces with legal guest workers.

Most farmers apparently believe that the supply of farm labor inside U.S. borders is not responsive to higher wages. Instead, they focus on improving the quality of life on the job and offering financial incentives to increase retention. In the past, many employers used sticks to reduce turnover, refusing to rehire seasonal workers who quit before the end of the season. Tighter labor markets make it harder for farmers to enforce such no-rehire rules, prompting more to offer bonuses that can add 5 to 10 percent to earnings for workers who stay through the season.

Stretching today's work force means raising productivity. And a lot of the potential for higher productivity is in the low-tech details. Workers harvesting by hand spend much of their time carrying produce down ladders to bins or to the end of rows to receive credit for their work. Smaller trees would mean fewer ladders and faster picking; hydraulic platforms would eliminate the need to fill heavy bags of fruit while standing on ladders. Slow-moving conveyor belts that travel ahead of workers harvesting berries, broccoli and other fruits and vegetables reduce carrying, making workers more productive and harvesting jobs more practical for older workers and women.

Under the Bracero "guest worker" program that lasted from World War II until 1964, most fruits and vegetables were packed in 50- to 60-pound field boxes, lifted by hand into trucks and taken to packing sheds. When there were fewer workers in the 1960s, incentives were created to switch to bulk bins that hold 1,000 pounds of apples or oranges along with forklifts to move the bins.

Note, too, that trees and field plants have been designed for maximum yields, not worker productivity. But dwarf trees, tall-stalk broccoli that requires less bending to cut and tabletop production of strawberries (already found in some European countries) could stretch the productivity of a smaller work force.

There is also the option of flat-out replacement of workers with machines. For most of the past century, the most labor-intensive farm activity in North America involved over 50,000 workers who harvested raisin grapes around the city of Fresno. Over a six-week period in August and September, workers cut bunches of green grapes and laid them in 25-pound batches on paper trays to dry into raisins in the sun, earning about a penny a pound.

Now, harvesting raisins is a race between sugar accumulation and spoilage. Allowing grapes to stay on the vine increases quality but raises the risk that September rains will damage the drying raisins. Raisins have traditionally been made from Thompson seedless grapes, but varieties such as Selma Pete achieve optimal sugar earlier in August, so that the canes holding bunches of green grapes can be cut and the grapes dried partially or fully into raisins while they are on the vine. A harvesting machine using rotating fingers knocks the partially dried raisins onto a continuous paper tray in the vineyard. Or the machine harvest is delayed until it can remove fully dried-on-the-vine raisins.

Two major factors are slowing raisin mechanization: farm structure and international trade. Most growers are over 60, have fully paid for their 20-40 acre vineyards, and are reluctant to make upfront investments in machinery when China, Iran and Turkey can already produce raisins cheaper.

The fourth adjustment strategy is to supplement the current work force with H-2A guest workers. The H-2A program was created in 1952 and was used primarily by sugar cane growers in Florida and apple growers along the East Coast until the mid-1990s. North Carolina tobacco farmers became the largest users after a group of retired government officials created an association that, for a fee, recruits workers in Mexico, brings them to the state and deploys them to farmers. This turnkey (and reliably available) H-2A labor force proved very attractive to farmers, especially as the workers gained experience by returning year after year.

Receiving government certification to employ H-2A guest workers requires employers to satisfy three major criteria. First, farmers must try to recruit Americans and provide good reasons why those who applied were not hired. Many farmers do not want to hire legal U.S. residents, since each one who promises to work blocks the admission of an H-2A guest worker. Indeed, farmers who are convinced that most U.S. workers will not remain for the entire season often discourage local workers from applying.

Second, farmers must provide free housing to H-2A guest workers and out-of-area U.S. workers. Housing is a special concern in California, where most labor-intensive agriculture is close to cities with shortages of affordable housing and restrictions on building more. During the Bracero era, most farm workers were housed on the farms where they worked, which meant low (or no) housing costs and no commute to work. A combination of tougher housing regulations and union hostility prompted most farmers to eliminate this housing in the 1970s, and today there is often community opposition to creating it – as demonstrated by an April 2016 arson fire in the town of Nipomo that destroyed dwellings meant to accommodate over 100 H-2A guest workers.

Third, the law requires that H-2A guest workers do not "adversely affect" American workers. The government enforces this requirement by setting a super-minimum wage called the Adverse-Effect Wage Rate, which was $11.89 an hour in California in 2016 – $1.89 more than the state's regular minimum of $10 an hour.

Now, all workers must be paid the AEWR, but farmers do not have to retain workers who are unable to pick fast enough to earn it. So the effect is to weed out slower pickers. Farmers selecting from a vast pool of eager foreign workers are more likely to find ones who can satisfy productivity requirements than if they recruit from the relatively small pool of U.S. workers willing to fill seasonal farm jobs.

Immigration Uncertainty

California's seasonal farm labor market has been a revolving door for the past century, attracting newcomers who stay in the seasonal farm work force a decade or two before moving to non-farm jobs. Their American-raised children shun farm work. Nothing new here: California history is a story of waves of newcomers – Chinese, Japanese, Filipinos, Dust Bowl Arkies and Okies, Mexicans – who passed through the farm labor market in a generation or less.

In the early 1980s, farm labor unions were weakening and wages were falling as the share of unauthorized workers rose toward 20 percent. A compromise included in the Immigration Reform and Control Act of 1986 aimed to reverse falling farm wages by legalizing unauthorized farm workers already employed, but imposing sanctions on employers who knowingly hired additional ones. Farm labor costs were expected to increase, as farmers raised wages to retain newly legalized workers or built housing to hire legal H-2A guest workers.

But the 1986 act backfired: it led to more rather than less illegal immigration. Farm wages fell in the 1990s and unauthorized workers spread to all commodities and states, while H-2A guest worker admissions dropped below 20,000. As the share of unauthorized workers rose toward 50 percent in the mid-1990s, farmers asserted that there was a shortage of legal U.S. workers to harvest their crops. They claimed agriculture needed an "E-Z guest worker" alternative that cut through the red tape, but in the face of union opposition, Congress declined to act.

The election in 2000 of presidents Fox in Mexico and Bush in the United States spurred farm employers and worker advocates to propose the Agricultural Job Opportunities, Benefits and Security Act (AgJOBS), another effort to legalize unauthorized farm workers and make it easier to hire guest workers. But, once again, the farmers' lobbying efforts failed: AgJOBS was never enacted.

For a while, farmers thought help was on the way from an unexpected direction. In November 2014, President Obama issued an executive order creating the Deferred Action for Parents of Americans and Lawful Permanent Residents (DAPA) program. If implemented, DAPA would have legalized four million unauthorized parents with legal U.S. children, including up to 500,000 farm workers, giving them renewable three-year work permits. But Texas and 25 other states sued to block implementation, arguing that DAPA was an unconstitutional overreach of executive power. Conservative federal judges agreed, and, on a 4-4 vote in June 2016, the Supreme Court let stand lower court injunctions blocking DAPA. This effectively ended the efforts of the Obama administration to ease the legal plight of unauthorized foreigners – and indirectly, stabilize farm labor supply.

What Next?

Neither drought nor labor shortages have deterred farmers from planting more labor-intensive commodities, which explains why farm sales and farm employment have been rising. Farmers complain of labor shortages that force them to leave food in the field. But this is a common cost of doing business: farmers regularly leave crops unharvested because of low prices and/or poor quality. Rising farm sales suggest that any crop losses from labor shortages are very localized.

AgJOBS would have locked the status quo in place, leaving agriculture with low-skilled and foreign-born workers. The absence of immigration reform has forced market adjustments – various measures to increase labor supply (domestic and foreign), to increase labor productivity and to mechanize more operations. And a continuing political stalemate over comprehensive immigration reform would accelerate this process.

In the short term, the dominant response has been expansion of the H-2A program, which has doubled in size nationally since 2007 and quadrupled in California. Over 10 percent of long-season jobs on crop farms nationwide and 3 percent of California crop jobs are now filled by H-2A workers, almost all from Mexico. If all H-2A workers were allowed to remain in the country for three years, the source of farm workers could well shift from Mexico to Asia. That would bring California back to the future, as when Chinese workers dominated the California harvests in the 1880s.

Plainly, this is a period of great uncertainty for agriculture in California – uncertainty largely driven by forces (political and natural) beyond the state's control. What is emerging, though, is a leaner, less tradition-bound industry that depends less on protective economic regulation and more on market forces.

That is good news for American consumers who rely heavily on California for a host of farm products. And it is good news for those of us who want the hundreds of rural communities dependent on California agriculture to avoid the sort of rapid dislocation that devastated Rust Belt manufacturing in the 1980s and 1990s.