Historical perspective
Provocative style
Psychological depth
Imagine two investors. The first promises you an average return of €75 – but with such volatility that you might end up with €200 or nothing at all. The second offers a modest €30, but almost no risk. Which one do you choose? The answer depends not on math, but on whether you can afford to lose. That question – disguised in the robes of academic respectability – lies at the heart of a recent study on how agricultural subsidies are distributed in Europe.
Money, as I like to remind people, is a collective hallucination. But subsidies are a hallucination squared: we believe the state knows whom to pay, how much, and why – so that fields stay fertile, farmers stay afloat, and society prospers. Reality, as usual, has its own sense of irony.
The Old Game with New Data
Economists have long mastered the art of optimization: give the medicine to those who’ll benefit most, award the grant to the firm with the greatest potential. It’s called «learning optimal policy» – a fancy term for a simple idea: look at the past to choose better for the future. The catch is that the classical model assumes a risk-neutral planner. He cares only about averages and ignores variance. A farmer might make €1,000 or €100 – as long as the average comes to €550, the planner is satisfied.
But imagine you’re not an abstract planner with a calculator, but a minister of agriculture before an election – or the head of a family farm with three kids waiting for the harvest to pay off the loan. Suddenly the average stops being comforting. What matters is not «how much on average», but «how safely can we avoid disaster?»
Enter a half-forgotten principle from the English economist Arthur Roy, formulated back in 1952: «Safety First.» Instead of chasing the highest average return, Roy proposed minimizing the chance that income falls below a critical level. In plain terms: better to secure a subsistence floor than gamble for a jackpot.
This logic sounds familiar to anyone who’s heard of the Sharpe ratio – the relationship between return and risk. The higher it is, the better: you earn more per unit of anxiety. In finance, that’s a trader’s tool; in social policy, it’s a matter of survival for entire regions.
The Formula That Says: «Not All That Glitters Is Gold»
A group of researchers decided to apply Roy’s principle to the problem of subsidy allocation. Instead of maximizing the average income of farmers, they aimed to maximize the probability that income stays above a socially acceptable minimum. The math is elegant yet intuitive: if outcomes are normally distributed (a rare luxury in economics), the optimal solution boils down to choosing the option with the best mean-to-standard-deviation ratio.
In other words: don’t chase the highest average yield if it comes with wild swings. Choose stability.
Sounds trivial? Perhaps. But the entire history of economic crises is a chronicle of people ignoring that triviality. From the tulip mania of the 17th century to the subprime mortgage meltdown of 2008 – each time we forgot that a high average return with high volatility means half the players end up at the bottom.
Two Strategies, Two Worlds
To illustrate the difference, the authors ran a simple simulation: two subsidy schemes. The first gives an average of €30 with low variability (a standard deviation of €10). The second offers €75, but with massive risk (a deviation of €65).
A risk-neutral planner would pick the second one – €75 beats €30. Pure logic. But a planner who worries some farmers might go bankrupt will think differently: the ratio of mean to risk is 3 for the first option (30/10) and about 1.15 for the second (75/65). The cautious strategy wins.
And what did the simulation show? On average, the aggressive strategy yields more. But in the worst-case scenarios – which do happen – farmers end up worse off than if they’d received no subsidies at all. The cautious strategy brings lower averages but almost never dips below the critical threshold.
That’s the social cost of uncertainty: to avoid catastrophe, we sacrifice part of our potential gains. Like an insurance policy – you pay the premium, hoping you’ll never need the payout.
Reality Check: Italian Farmers and Brussels Money
Theory is one thing; how does it fare in the real world? The researchers drew on the Italian Farm Accountancy Data Network – a massive dataset covering 9,336 farms over thirteen years (2010–2022). Their focus: the first «pillar» of the EU’s Common Agricultural Policy – direct payments to farmers, which make up more than half the CAP budget.
These subsidies are mostly allocated by historical entitlements and land size. The logic is simple: more hectares, more money. The system runs smoothly, but there’s a catch – it ignores how efficiently that land is used, how much it produces, and how risky the activity is.
The researchers asked: what if subsidies were distributed not by acreage but by an optimal rule reflecting each farm’s characteristics? They considered machinery capacity, labor costs, fixed and variable expenses, participation in other programs. Performance was measured by net agricultural output per hectare – roughly, how much real value a farm creates from each unit of land.
Subsidies were divided into three levels: low, medium, and high. The planner’s task: choose the optimal support level for each farm.
What the Numbers Showed: The Triumph of Caution
The results were striking. Three scenarios were compared: total risk-neutrality, moderate caution, and strong loss aversion.
In the first scenario (risk-neutral), the actual subsidy distribution produced an average welfare index of 11.68. The optimized model raised it to 13.63 – a 17% gain. That’s no small feat when billions of euros and tens of thousands of livelihoods are at stake. Curiously, real bureaucratic decisions matched the optimal ones in only 31% of cases – meaning two-thirds of the time, funds went astray.
In the second scenario (moderate caution), actual welfare fell to 7.80, while the optimal policy reached 11.16 – a 43% improvement. The planner sacrifices some average welfare for less volatility, yet even then the optimized policy beats reality hands-down.
In the third scenario (maximum caution), the actual value was 6.34 versus an optimal 11.12 – the gap widens, but the overlap of real and optimal decisions remains at 31%.
The conclusion writes itself: the current subsidy system is systematically inefficient. It maximizes neither average income nor stability. It simply perpetuates an old structure – a historical compromise between lobbyists, bureaucrats, and politicians.
Yet there’s a deeper insight: the more cautious we are, the lower the overall welfare. That’s the social price of insurance. We pay to ensure no one falls through the cracks – but the price is a lower average standard of living.
The Stability Paradox: Why We Fear Winning
There’s something deeply human in that trade-off. We’d rather forgo the chance to grow rich than risk losing everything. That’s not irrational – it’s evolutionary wisdom. In a hunter-gatherer world, one failed season could mean extinction. Better to gather fewer berries for sure than chase a mammoth and starve.
But in modern economies, that instinct turns against us. Fear of volatility keeps governments clinging to inefficient yet predictable systems. Farmers get subsidies not because they perform well, but because they always have. Banks get bailouts not because they create value, but because their collapse is scarier than their mediocrity.
Here’s the irony: in trying to avoid risk, we build systems that guarantee mediocrity. Like an insurance firm that covers everyone – cautious drivers and reckless ones alike. Everyone pays more, yet driving doesn’t get safer.
A Tool for the Real World
The researchers offer not an abstract theory but a working instrument. Their method can adapt to many contexts: add budget limits (money is always scarce), fairness concerns (why do rich farms get more?), or dynamic effects (how today’s choices shape tomorrow’s outcomes).
You can go further still – using not just the mean and standard deviation, but more sophisticated risk metrics like quantiles or Conditional Value-at-Risk (CVaR). That way, you protect against not average volatility but the «tail» events – those rare catastrophes that break everything.
In the end, it all comes down to one question: are we optimizing for higher income – or for peace of mind?
A Mirror for Policymakers
Step outside agriculture, and the same approach applies anywhere uncertainty meets public decision-making: education grants, health insurance, pension systems. Everywhere we distribute scarce resources among people with diverse profiles and unpredictable outcomes.
The classical approach – give everyone equally, or more to those who «on average» deliver greater returns – misses the core truth: life is not lived in averages. Real people live in specific realizations of randomness. If yours happens to land in the «lower tail», the statement «on average it’s fine» offers no comfort.
Risk-sensitive policy makes visible what formulas hide: the choice between efficiency and insurance. Do we want to maximize the pie – or ensure everyone gets at least a slice? That’s not a technical question. It’s moral and political.
And here’s the kicker: the current subsidy system delivers the worst of both worlds. It neither maximizes average income (too tied to the past) nor minimizes risk (blind to volatility). It merely persists – a monument to mid-20th-century compromise.
The Price of the Control Illusion
Why don’t politicians reform it? Because doing so means admitting the old rules were flawed – a form of political suicide. It’s easier to say «We support traditional agriculture» than to explain why Dupont’s 500-hectare, robot-driven farm is more efficient than Lefevre’s smallholding, tilled by hand like his grandfather’s.
There’s another reason: the current system offers the illusion of control. A bureaucrat can open a registry, check land size, and calculate payments. Transparent. Predictable. The new method demands data, models, risk assessment – and the humility to accept that outcomes may surprise us.
But the illusion of control is yet another collective hallucination. We think we manage the system – yet it manages us, repeating patterns long devoid of meaning.
What Comes Next?
The study doesn’t claim to be revolutionary. It offers a tool – transparent, mathematically sound, practical. A tool for those brave enough to ask a simple question: can we allocate resources better than we do now?
The answer, judging by the Italian data, is yes. We can. The difference isn’t in single digits – it’s in tens of percent. Not a tweak, but a leap.
But – and here the irony of human nature returns – a tool alone changes nothing. It takes will to use it. And will requires faith: faith in data, in models, in the belief that the future can be improved without tearing everything down.
Money is a collective hallucination. But the politics of distributing it – that’s a hallucination cubed. We believe we know best, even when the data says otherwise. We claim to value stability over efficiency, yet build systems that deliver neither.
Perhaps it’s time to wake up. Or at least admit that we’re dreaming.
So the choice remains – not yours or mine, but of those who manage budgets and sign laws. Chase the highest average income and risk catastrophe? Or choose modest stability and pay for it with a bit of growth?
There’s no «right» answer. Only the honesty to admit it is a choice – and that every decision carries a price, visible or hidden, but always real.
