Before a production cycle starts, work out how far the price can fall, how far the harvest can fall, and how far both can fall in the same year before the cycle stops covering its costs. The cycle can be a crop from planting to sale or a year of milk or beef. Each distance is the cycle’s room. On a cycle that expects 900,000 in revenue, in whatever currency you keep your books, and costs 720,000 to run, the price can fall 20 percent, the harvest 23.1 percent, and both together only 11.4 percent. The combined figure is about half the others, and it is the one to write down first. The shock resilience calculator works out all three.
Rurivia calculation from an example farm’s expected revenue and cost, run through the shock resilience calculator.
Then compare that room with the typical swing: how far a normal year’s price and harvest land from their average, in percent. On the example, eight years of records give 17.4 percent for price and 8.9 percent for yield. The calculator’s resilience index asks: if price and yield fall in the same year, each by the same fraction of its own typical swing, how large can that fraction get before the profit is gone? Here the answer is 0.84, less than one. The price down 14.6 percent (0.84 of 17.4) and the harvest down 7.5 percent (0.84 of 8.9) at once wipe out the whole margin, and each fall is smaller than one typical swing.
Many farms plan the other way. A price drop feels like a market problem and a drought like a weather problem, so each gets its own protection, and often one of them gets none. The year that hurts is the one when both arrive together.
Why is the room on yield larger than the room on price?
Because a smaller harvest also costs less to bring in, and a lower price does not.
When the price falls, revenue falls and the costs stay where they were. A bad price at harvest does not change what the seed, the fertilizer and the diesel already cost. So the room on price is the simplest figure: divide the profit by the revenue. In the example, 180,000 over 900,000 is 20 percent, and at a 20 percent price drop the cycle breaks even.
When the harvest falls, revenue falls, and so does every cost paid per unit harvested: custom harvesting charged by the bushel or the tonne, freight to the elevator or the port, drying, storage, the sales commission, a technology fee charged per unit delivered. None of it gets spent on grain that never came in. In the example, 120,000 of the 720,000 moves with the size of the harvest, and the other 600,000 stays put whatever comes off the field.
Work it through and the harvest can fall 23.1 percent before the cycle breaks even. At that point revenue, less the 120,000 of harvest-linked cost that shrinks in the same proportion, just covers the 600,000 that does not move. Whenever part of the cost moves with the harvest, the room on yield comes out larger than the room on price, on any cycle that closes with a profit.
The gap between 23.1 and 20 percent, about three percentage points, may look like a detail, but it tells you which fall wipes out the profit first. A farm that believes both rooms are the same size will treat a drought and a price collapse as equally dangerous. On any farm that pays anything by the bushel or the tonne, the price needs the smaller fall to get there.
A standard crop budget lists the cycle’s revenue and costs at the expected yield, and counts harvesting and freight as fixed amounts. That is fine for asking what the cycle should earn, which is how our crop budget article, a step-by-step guide to what a cycle should leave you, uses one. But work out the break-even from that budget and price and yield get the same room, because no cost shrinks with the harvest. To ask how far the cycle can fall, keep the two rooms apart.
Which cost lines fall with the harvest, and which are already spent?
This is a judgment call, and no calculator makes it for you. The split that matters is not the one your accountant uses.
Fixed against variable asks whether a cost changes with how much you plant. That is the wrong question here. Seed, fertilizer and crop protection are variable costs in any textbook, yet they are paid by the time anyone knows the crop failed, and a short harvest gives back none of them. For this calculation they sit with the costs that do not move, next to land rent and depreciation.
The question to ask of each cost is whether it depends on how much actually leaves the farm. Go down last cycle’s invoices and put each line on one side or the other:
Moves with the harvest: custom harvesting paid by the bushel or the tonne, freight to the elevator or the port, drying and cleaning charged by the load, storage paid per unit stored, the sales commission, any technology fee charged per unit delivered, and the packaging, twine or bags used only on what was harvested.
Already spent: seed, fertilizer, lime, crop protection, the fuel burned before harvest, land rent, machinery depreciation, wages, insurance, overhead, interest on the operating loan, and custom harvesting paid by the acre rather than by the bushel or the tonne.
Two lines usually cause an argument. Custom harvesting is the first, and its side depends on the contract, not on the work: paid by the acre it is already spent, paid by the tonne it moves. The second is your own machinery. Running the combine over a thin crop burns less fuel and wear than over a heavy one, but not proportionally less, because the machine still crosses the whole field. Put the part that scales with the tonne on the moving side, leave the rest where it is, and write down how you decided, because a year from now nobody will remember.
Getting this split wrong costs more in one direction than in the other. Leave the calculator’s line for harvest-linked cost blank and it treats no cost as moving: the room on yield drops back to 20 percent instead of 23.1, and the index to 0.80 instead of 0.84. That error is small and on the safe side, because it understates the room.
Overstating the moving share is the dangerous error. Declare all 720,000 as moving with the harvest and the calculator says the harvest could fall to zero and still break even, because in that arithmetic a crop that does not exist costs nothing to bring in. The result is right inside the formula and absurd on the farm, which is what any formula returns when it is fed a figure nobody checked.
What does twenty percent of room not tell you?
Whether twenty percent is a lot.
The Food and Agriculture Organization of the United Nations (FAO) calls these three figures switching values in its guidelines for agricultural investment projects, and warns that they measure how sensitive the result is and not how likely the fall is: “two dimensions should be considered: sensitivity and probability”.
Twenty percent of room means one thing on a dairy selling milk under a yearly contract and something else on a soybean farm selling at the spot price. The percentage is the same. What separates the two farms is how far their own prices and yields have moved in past years, and that is in their records, not in the budget.
So the probability side has to come from the farm’s own records. For the price series and for the yield series, work out the typical swing, which is the standard deviation divided by the average; the calculator does this for you from the list of yearly figures. Then ask how many of those swings fit inside the room before the profit hits zero. In the example, the price swings 17.4 percent and the yield 8.9 percent, and the combined room holds 0.84 of one swing in both at once. That number is the resilience index.
The calculator reads the index in three bands. Under one is fragile: an ordinary bad year on that farm wipes out the margin. Between one and two is tight: the loss sits inside what the records already show. Above two is roomy: it takes a year worse than anything in the records.
Before trusting that division, do the arithmetic that needs no method at all: count. In the example’s eight years of prices, five of the seven moves from one year to the next were falls, and three of those were bigger than 15 percent. Two of the eight years closed more than 11.4 percent below the series average. That count assumes nothing, and it is the first thing to write on the page.
A count has limits of its own. In eight years, the years when price and harvest fell together usually number zero or one, and a zero only means too little data. A count also ignores size: a year that went just past the room counts the same as one that went far past it. The index uses all eight years, keeps price and yield in a single reading, and measures how far away the loss is rather than how often it comes. Dividing the room by the swing is this article’s own choice, not a published standard, so keep the count beside the index on the page.
Where do both together take the farm down?
The combined room is roughly half of either single room, because revenue is price times harvest, so the two falls multiply. Lose a tenth of the price and a tenth of the harvest and revenue falls by 19 percent, since 0.9 times 0.9 is 0.81, while the costs that do not move stay whole and eat the difference. On a 20 percent margin with no cost moving with the harvest, an equal fall of 10.6 percent in both is enough to break even. With the example’s harvest-linked cost, it takes 11.4.
The usual objection is that price and yield do not fall together, because a short crop brings a high price. That holds for a country and often fails for a single farm. The Economic Research Service of the United States Department of Agriculture explains why: “Yield and price on a farm, for example, need not be related because the output of one farm does not noticeably affect market prices.”
The offset, which the same report calls a natural hedge, works for farms inside the region whose harvest sets the price. Outside it, “the natural hedge is much weaker, meaning that low yields and low prices (or conversely, high yields and high prices) are more likely to occur simultaneously”. The report adds that “Yield variability is higher at the farm level than at the State or national level.”
The two falls can also come from separate causes in the same year: in Central America, the 2012 and 2013 coffee crop met a fall in international prices and production losses to leaf rust at the same time. The regional agricultural council put the loss at 500 million dollars, a fifth of the harvest, according to a study by the United Nations Economic Commission for Latin America and the Caribbean (ECLAC).
So the combined room is the realistic case for a farm outside the center of its market, and the pessimistic case for one inside it.
You can check which applies to you in your own records. Line up the years and look at the weakest harvests. If the price was above average in those years, the offset works for you; if the price was low too, it does not. In the example, the two worst harvest years, with yields of 49 and 55 against an average of 59, were the two best price years, 115 and 130 against an average of 98.5, so the offset works there.
A farm with the same budget whose bad harvest years were also its bad price years has no offset to lean on, so its 0.84 is the reading to take as it stands. Where the offset works, as in the example, the real index sits somewhat above 0.84. Check which case is yours before treating the index as a verdict, and write the answer down.
Why must the swing come from your records, not from memory?
Because the swing people estimate from memory tends to come out too narrow, and the calculator cannot tell a remembered figure from a recorded one.
When the series is not at hand, the calculator accepts a typed percentage and warns you when you use it. The warning rests on a study published in the Revista de Economia e Sociologia Rural, a Brazilian journal of farm economics, in which Cruz Júnior and colleagues asked Brazilian maize farmers to spread their price expectations across ranges; state by state, in every case the standard deviation obtained from the questionnaires was lower than the historical one. It was one crop in one country, but the direction is a reason to use the record.
A narrow swing does not fail quietly: it flips the verdict. Type 12 percent for price and 8 for yield instead of the 17.4 and 8.9 in the example’s records, and the index rises from 0.84 to 1.12, from fragile to tight. Nothing on the farm changed. One figure came from the records and the other from memory.
Eight years is also a short record. Zampieri and colleagues, who built a published indicator of how steady crop production is, report that its sampling error stays under 30 percent only with at least thirty years of data. They add that “Larger sampling errors are associated to shorter time series and for crop systems displaying larger inter-annual fluctuations compared to the production mean.” Eight years gives a rough measure, still far better than none, and the roughness shows in the index instead of hiding.
If only one series can be rebuilt, start with price. Oliveira and colleagues, in the Revista de Política Agrícola, a Brazilian farm policy journal, studied three dairy systems in Piracanjuba, in the state of Goiás, and found that milk price and yield were the most relevant variables for the profitability indicators, and price proved more significant than yield; it covered one region and one activity, so check it against your own series.
What is this index not?
It is not a probability, and it is not the only resilience figure the same data can produce.
Run the example’s yield series through the indicator Zampieri and colleagues published and it scores high, because that series does not swing much. Their indicator measures steadiness and says nothing about where break-even sits, so a farm that loses money every year, steadily, could still score well on it. The index in this article measures the distance to break-even, counted in swings.
Four more things the index does not tell you.
It does not give the odds. Turning 0.84 of a swing into a chance of loss would mean assuming a shape for how prices and yields are spread, and the calculator assumes none. Under one means the distance to a loss is shorter than one ordinary swing, and nothing more precise than that.
It does not know whether your two series move together. The calculation pushes price and yield down side by side, each at its own scale, which is the right worst case for a farm outside its market’s center and a pessimistic one inside it. Where a real offset exists, the true index is above the one printed.
It measures the fall from your expected price, not from your historical average. If the expected price is above what your records average, part of the room is spent before the first bad day. Build the revenue on the price you expect, not the one you hope for.
And it starts from a budget, which is a set of numbers that have not happened yet.
What changes at each band?
There are two levers, and knowing which one a decision pulls matters more than the decision itself. You either widen the room or narrow the swing. Widening the room means a bigger margin: lower cost, higher yield, a better price, or less overhead charged to this cycle. Narrowing the swing means less variation in your own numbers: a contract, a hedge, irrigation, spread planting dates, more than one source of revenue. Farms often reach for the second lever when the cheaper fix is in the first.
The two levers pull against each other. Every protection that narrows the swing is paid out of the margin, and the margin is the room. Suppose that locking in the price on half the volume halves the price swing and costs 2 percent of expected revenue: the room shrinks, the swing shrinks more, and the example’s index rises from 0.84 to 1.19, from fragile to tight. The same protection at 8 percent of revenue gives 0.83, below where it started, so the farm paid to be worse off. In this example the break point is near 8 percent of revenue, or 40 percent of the margin. These are this article’s own calculations on the example farm.
Under one (fragile) is not a warning of disaster. It means an ordinary year, of a size already in your records, ends this cycle at zero or below. Decide before planting, not after, and put a price on each lever. Widening: what is the largest cost line, and what would it take to move it enough to raise the margin by a third, which is about what takes this example’s index above one. Narrowing: what would it cost to lock in part of the price, and is that cost smaller than the margin it protects.
The price and hedging scenario calculator puts a number on the second question, and the forward pricing calculator on what certainty costs; two articles walk through each one, what price you keep after hedging and is locking the price now worth it. A fragile index with no answer on either lever is a decision to carry the risk, which is sound only when somebody made it on purpose and wrote it down.
Between one and two (tight) means the loss sits inside what your records already show, so a year that reaches it is likely to come. A written trigger pays off here, and it comes from the combined room, not the room on price. In the example, that means acting when the price is 11.4 percent below the expected price rather than waiting for 20, because the 20 assumes the harvest comes in as planned. Write the price, the action and the name of the person who acts.
Above two (roomy), the question moves from surviving the year to keeping the room. The room follows the margin closely, and with swings like the example’s, roomy is rare: on the same farm, a 10 percent margin gives an index of 0.41, a 30 percent margin 1.30, and only a margin near 45 percent reaches two. A roomy cycle can turn tight after one cost increase, so the index is a reading with a date on it, not a trait of the farm.
The index also helps with a written list of the farm’s risks. If you keep one and rate each risk from one to five on how much damage it would do (our farm risk list article shows how to build it), a price drop and a drought usually sit on two separate lines, often with the same rating. The combined room puts money on what happens when both arrive: in the example, a 15 percent price fall alone leaves the cycle 45,000 ahead, and the same fall with a harvest 10 percent short leaves it 19,500 in the red. Adding the short harvest moves the result by 64,500, a gap that two equal ratings on separate lines would hide.
Where to start
An afternoon, with last cycle’s invoices, the price record and the yield records on the table. What comes out is one dated page, with more on it than the calculator’s inputs.
More articles on measuring what the farm is exposed to are in sustainability and risk, and farm management sets out how a reading like this one fits into running the whole farm.