METHODOLOGY
How the numbers are made.
Every figure on FarmHive comes from a published source or from data growers chose to share, through calculations set out on this page. They are benchmarks: they show where a crop or a farm stands against a reference. They are not forecasts of yields or prices.
1. Sources
- Ministry of Agriculture crop budgets (Shaham economic models): 234 per-dunam budgets. Field crops 2024 (22), vegetables 2023 (101), deciduous orchards 2023 (57), citrus 2026 (24), subtropical 2023 (11), and 19 orchard crops with no newer budget from the 2020 orchard booklet (cherry, olive, pomegranate, dates, figs and others). Each budget lists every input with its quantity, unit and price, the products sold with yield and price, and for orchards the costs and yields of every year of the orchard's life.
- Tzrifin wholesale market prices (prices.moag.gov.il): daily prices of local produce by crop, variety and grade, from November 2015, in nominal shekels.
- MIGAL orchard data: yields per plot and season for apple, pear, avocado, kiwifruit and cherry, and water per dunam for part of the apple and kiwi plots. The raw files stay with MIGAL; FarmHive publishes only figures for groups of plots (section 11).
- What growers share, by choice: season records and farm-gate prices (section 17), and protocols written by advisors and growers.
2. From crop budget to protocol
A protocol is a crop budget written in one common format, so that the Ministry's budgets, advisors' protocols and growers' own versions can be read, compared and adapted in the same way.
- Every line is read from the Ministry worksheet: the input, its group (labour, water, fertilizers, plant protection, machinery, materials, services, insurance, overheads, interest), quantity, unit, price and cost per dunam, and the products sold.
- Every protocol reconciles to its source. Each cost group and the total cost must equal the Ministry worksheet; all 234 do. (In two vegetable budgets, VG-076 and VG-091, the worksheet's fertilizer subtotal differs from its own lines; the totals still match.)
- Inputs get a standard name in English and a price driver. The same input is written in many ways across budgets. A vocabulary groups them into standard inputs and into 699 price drivers, the prices you can change (for example hired labour per work-day, fresh water per m³, one fertilizer per kg). All work-day labour maps to two drivers (hired and seasonal; permanent and management) and water to four (fresh, mixed, reclaimed, marginal). The first-pass vocabulary is under review by crop advisors.
- Quantities that give the cost. In about one line in seven the budget's quantity column counts something else: for sprays often the number of applications, for percentage costs the base. There the quantity is worked out as cost ÷ price, and the budget's own figure is kept and shown with a note. Where it counts applications, the builder offers rate per application × applications.
- Two levels. Economic protocols say what is used and what it costs. Rules protocols also say when each job is done and the decision rules behind it (what to watch for, the threshold, the response). Today five protocols have Rules from official agronomic sources: apple Cripps Pink and Gala, pear Spadona, avocado Hass and Ettinger (Shaham growing manuals, the 2025 Galilee–Golan bulletins and the Ministry avocado work instructions). Timings proposed by FarmHive are marked "to confirm with an agronomist".
A protocol written by an advisor or a grower is checked for format and arithmetic (the same checker runs on the website and in the build). Its agronomy is the author's; it shows as "Not reviewed" until FarmHive reviews it.
3. Direct costs and the full budget
Comparisons use direct production costs by default: inputs, labour, water, machinery and services. Insurance, financing (working-capital interest, capital recovery) and overheads (vehicle, office, municipal tax, land lease, fees, contingency) stay in the data and are added with the box "Include insurance, interest and overheads". They differ from farm to farm and say little about the crop itself.
Revenue is each product's yield × price in the budget. Orchards compare on the mature, bearing year unless the whole-life view is chosen (section 7).
4. Budget regions
FarmHive uses 15 regions in four areas. A budget belongs to the regions its source names: the worksheet's region field, else a place in its title. An area word (north, south, Negev, Galilee) covers every region in that area, and "whole country except …" keeps the exception. Budgets that name no region are national. A budget is offered in a region when it names the region, or when it is national and national budgets are included and its source does not exclude the region.
A region says where the Ministry built a budget. It is not a claim that the crop suits the region.
5. Crop profitability
Every budget with revenue is computed at the chosen settings and ranked. Three measures:
- Profit per dunam, as in section 3.
- Profit per m³ of irrigation water = profit per dunam ÷ the budget's irrigation water per dunam (the sum of its water lines). Rain-fed budgets have no irrigation water; they are left out of this measure and counted below the chart.
- Profit per labour day = profit per dunam ÷ the budget's work-days per dunam (its labour lines in work-days). Harvest labour paid per kilo in some vegetable budgets is not in work-days and so not counted.
The same arithmetic runs in the browser and in the Crop Choice workbook; an automatic test requires them to agree to 0.05 ₪ for every budget.
6. Bad, typical and good price years
A typical year uses the budget's own sale price. For a bad or good year the price is multiplied by a factor from the crop's price history at Tzrifin, 2015–2025:
- Each market line (variety and grade) is turned into a yearly average, relative to that line's own median; the lines are averaged (in logs) into one yearly price index for the crop.
- Volatility σ is the standard deviation of the year-to-year change in the log of the index. Using changes rather than levels keeps steady inflation out.
- Bad year = price × e−1.2816σ; good year = price × e+1.2816σ. This is roughly a one-in-ten bad or good year.
Crops with no market series (contract and processing crops) use an assumed σ of 0.19, the median of the series.
7. Orchards over their whole life
An orchard budget gives costs and revenue for every year, from planting through establishment to full bearing. The mature year shows what a bearing orchard earns but hides the years before it. The whole-life view turns the life into one figure per year:
Payback is the first year in which the running balance turns positive. For example, Cripps Pink earns 2,556 ₪ per dunam in the mature year on the full budget but −479 ₪ a year over its whole life, with payback in year 17. Payback and the establishment deficit are on the full budget; with direct costs only, the overhead share of the mature-year budget is taken off each year (exact for mature years, an approximation for establishment years).
Expected yield by age = the mature yield × the budget's yield share for that age. It is used to compare a young orchard with its budget.
8. Input prices
For each price driver: the typical price (the median of the prices the budgets use), the range across budgets, and how many budgets use it. Next to it, prices from your own season records, read in your browser: a changed line's actual price, otherwise the price in your version of the protocol. Lines left at the Ministry price or skipped do not count. A line's price is converted to the driver's unit with the line's factor; where several lines use one driver, the median is shown.
9. Market price spread
For 89 market series (a crop, or a variety where the market quotes it separately) from the Tzrifin daily prices:
- Each year: the 10th, 25th, 50th, 75th and 90th percentile of the daily prices over all grades. The middle half is the 25th to 75th; the middle 80% the 10th to 90th. The first year (2015, from 1 November) and the current year are marked partial.
- Each month over the last five full years, and each grade over the last three.
- Volatility: the year-to-year change in the yearly median, as in section 6.
- The budgets on the line: 145 Ministry budgets are matched to a series. Each budget's price per kg (per head for lettuce) is placed among the market days of the budget's own year and of the last three years: "below 9 in 10 market days" means under the 10th percentile.
Series the market stopped quoting (fewer than 60 prices in the last three years) are left out and listed. Prices are nominal. The matching of budgets to market series is a draft and is being checked.
10. Farm-gate share
A crop budget prices produce at the farm or packhouse gate; Tzrifin is a wholesale market. The farm-gate share compares them in the same year, so inflation does not enter:
If the budget's year has fewer than 20 market prices, the last three full years are used, and the page says so. 100% means the farm gets the wholesale median. Shares over 100% usually point to a different grade, season or market and are flagged. For growers' own prices the share is the median, over farms, of each farm-year's price ÷ that year's Tzrifin median, shown once at least five farms have shared (section 17).
11. Orchard comparisons from MIGAL data
- Groups: plots are grouped by crop and variety and matched to the Ministry budget for that variety (the traditional, standard-density budget; all cherry varieties to the one cherry budget). Varieties with no budget of their own form their own groups and are compared with each other. Plots are never compared across varieties or production systems.
- Headline yields: mature plots (6 years and older) in the ten seasons 2013–2022, as the 10th, 25th, 50th, 75th and 90th percentile of the plot-season yields (t per dunam), next to the budget's yield.
- Yield by age of all plots, against the budget's expected yield for that age (section 7), and the median by season.
- Fresh or reclaimed water: where the file records the water source (apple and avocado; kiwifruit records the water quality) and both groups reach the minimum, the median mature yield on fresh and on reclaimed water. A third source in MIGAL's files, Zemer water, is kept out until its kind is confirmed.
- Privacy minimum: a figure is published only when it rests on plots from at least 5 settlements (since 9 November 2026 every MIGAL file names the settlement; the older apple file did not, and used a minimum of 10 plots). Smaller cells (an age, a year) are left out. No names, settlements or plots leave the raw files. Rows MIGAL marks as exceptional are left out.
- Counts on the site are in plots and plot-seasons; the settlement count is used only for the minimum. A plot is one block of one variety, planted in one year, so the plots of a crop's varieties add up to the crop's plots. A crop's Plots counts plots of every age and season; each median uses only the mature plots in the comparison window.
12. Water use
For crops whose MIGAL files record water (apple, from the Naftali hills file, and kiwifruit): the average m³ per dunam per season over all plots with water records in 2013–2022, the middle 80% range (10th to 90th percentile), and m³ per tonne of fruit. Your own figure is placed on the same bar; beyond the 10th or 90th percentile the page says "less / more than 90%" rather than a precise rank.
In My Farm, a season's water is the sum of its water lines: as recorded, or the protocol's amount where unchanged. Fresh versus reclaimed water is not shown: the reclaimed-water plots come from too few settlements to meet the minimum.
13. My Farm: your prices and your plan
Each plot follows a protocol: the Ministry's as it is, or your own version with your quantities, prices, yields and sale prices. Prices are applied in this order:
Your fresh and reclaimed (or brackish) water prices replace the protocol's water prices in every plan. The comparison with the Ministry protocol always uses the Ministry's prices. For a young orchard the plan is scaled to its age with the budget's yield shares.
Everything you type stays in your browser.
14. Season against plan
A season record holds, for each line of the plan, whether it went as planned, changed or was skipped, the actual quantity and price, and a reason; and for each product the actual yield and sale price. The difference in profit from the plan is split into parts that add up exactly:
- Yield = (actual yield − planned yield) × planned price, for each product;
- Sale price = (actual price − planned price) × actual yield, for each product;
- Costs, one part per reason = the sum of (planned − actual cost) over the lines given that reason ("Reason not given" when none is chosen).
15. Decide: ranking crops for a plot
For a plot, every protocol offered in your region is computed at your prices and water prices, then ranked by the measure you choose: profit, profit per m³ of water, profit per labour day, or profit in a bad price year. The result is shown per dunam and for the whole plot. It ranks budgets; it does not say a crop will grow well on your land.
16. Simulate: other prices, wages and yields
The simulator computes every crop budget twice, with the same region, price year, cost scope and orchard basis: once at the budget's original prices (Original) and once with the changes you set (Simulated). The difference is the effect of your changes alone.
- A % change for a price group (water, labour, fertilizers, plant protection, machinery, seed and plants, harvest and packing, other) multiplies every line of that group in every budget, each from the budget's own price: water +50% turns 1.98 ₪ per m³ into 2.97 in one budget and 1.40 into 2.10 in another.
- An exact price for fresh water, reclaimed or brackish water, hired labour or permanent labour replaces the budget's price for that input in every budget, as in My Farm, and overrides the % for those lines.
- Sale prices multiply the revenue of every product. Yield multiplies the quantity sold; costs, including harvest, packing and marketing, stay as in the budget, so a smaller harvest is not partly made up by lower picking costs.
- The price year (section 6), the cost scope (section 3) and the orchard basis (section 7) apply to both columns.
For one crop the page also shows:
- What moves its profit most: the change in profit per dunam if each lever rose by 10% from the simulated level: 10% of revenue for sale prices and for yield, and 10% of the group's cost for each price group.
- Where it breaks even, one change at a time with everything else as simulated: sale prices or yield could fall by profit ÷ revenue; water could cost up to (water cost + profit) ÷ m³ of water; labour up to (labour cost + profit) ÷ work-days. For orchards these use the mature year.
A scenario is kept in the page address, so a link reproduces it. Nothing is sent anywhere. A simulation shows what the budgets give under the conditions you set; it is not a forecast of those conditions.
18. Reading form 1220
An uploaded form is read in your browser and not sent anywhere. On the official layout the cells are found by position, and the personal areas (§1, §2, block and parcel numbers, owners and tenants, §9 suppliers, §10 and the signature) are dropped before anything else reads the text. The crop lines of §3, §4 and §6 and the sold quantities of §8.3 are kept. Each line is matched to a Ministry budget by its Hebrew crop name (the longest name first), then variety, then the rain-fed or irrigated column. Every line is shown for checking before it is used.
19. How the calculations are checked
- Every protocol reconciles to its Ministry worksheet, by cost group and in total.
- The website's engine and the Crop Choice workbook are run on the same settings and changed prices and must agree to 0.05 ₪.
- The site is walked in a browser by automatic tests that recompute the figures on each page independently from the published data: farm totals, season splits, rankings, market and MIGAL figures.
- Market figures were checked against an independent recalculation from the raw price file.
20. Known limits
- Budgets are for the year and region they were built for; the 2020 orchard budgets are older than the rest.
- Some budgets are national or from another region and stand in for the local one (for example wheat, cotton and silage).
- Harvest and marketing costs do not change with the yield or price settings.
- Vegetable harvest labour paid per kilo is not counted in work-days.
- Five 2020 orchard budgets carry no pest-control cost (their pest tables have no prices); the avocado budgets likewise.
- The matching of budgets to market series and the English input names are drafts under review.
- MIGAL figures describe the plots in MIGAL's files, mostly in the north; they are not a sample of every orchard in Israel.
Questions or corrections: adama@migal.org.il