Lesson preview · Models and data
Models, assumptions and "all else equal"
~7 min · Free to read
Imagine Kai runs an ice-cream stall on a beach. Tomorrow will be hotter than today. How many cones should he get ready?
Lots of things change his sales: the weather, his price, a rival van, school holidays. He can’t weigh them all at once, so he keeps the one that matters most: on hotter days, he sells more cones. That simple rule is a model.
Key Term
Models and assumptions
An economic model is a simplified picture of how one part of the economy works. Like a map that shows roads but not every tree, it leaves most things out on purpose.
A model’s assumptions are what it takes for granted, such as “Kai’s price stays the same”. They are rarely perfectly true. Most economists judge a model mainly by how well it predicts for its question. An assumption causes trouble when what it sets aside matters for that question.
Key Term
All else equal
To see what one cause does, change only that cause and keep the rest fixed, as in an experiment. Economists call this ceteris paribus, Latin for “other things being equal”, or all else equal.
“All else equal, Kai sells more cones on a hotter day” is a narrow claim. It says what heat does on its own, not that nothing else will change tomorrow.
Worked Example
Kai’s records show that on otherwise similar days between 18°C and 30°C, he sells about 20 more cones for each extra degree. On Saturday it was 22°C, and he sold 200 cones at 3 dollars each. Sunday’s forecast is 27°C. Use his model to predict Sunday’s sales.
Check yourself · no marks
What did we hold constant? Name three things Kai assumed would stay the same. If one of them changed, how would his real sales compare with 300?
Commit to one first. Guessing and being wrong beats reading the answer cold.
On hot days, Kai sells more ice cream, and more people get sunburnt. Ice cream doesn’t cause sunburn: the sunshine causes both. When two things rise together, a third factor may drive both, or the link may be pure chance. To show a cause, check that the link holds when other things are equal.
Even a careful model can fail when something it takes as given changes. In December 2021, the European Central Bank published a forecast that prices in the countries using the euro would rise by 3.2% in 2022. It assumed energy prices would fall, as markets then expected. But they were already higher than expected in early 2022. Then Russia invaded Ukraine in February, and they climbed far higher. Across 2022, prices were on average 8.4% higher than in 2021. In spring 2022, the bank’s economists checked its recent misses. Wrong assumptions, mostly about energy prices, explained about three-quarters of them.
Trap · Common Misconception
"Unrealistic assumptions make a model useless"
Kai’s model leaves out the wind, rival sellers and much else. It is still useful: it predicts well on ordinary summer days.
A model can go wrong in two ways. Something it takes as given may turn out differently, as energy prices did in 2022. Or it may be used far from the cases it was built on: at 40°C, people may stay off the beach.
Key Term
Positive and normative statements
A positive statement says what is, was or will be, such as “If Kai charges 4 dollars, he’ll sell fewer cones.” Evidence can check it, even if it turns out wrong.
A normative statement says what should be, such as “Kai should charge less, so every child can afford a cone.” It rests on values, so data alone can’t settle it.
Models answer positive questions. What should happen also takes a value judgement.
Tip
"Should" is a clue, not a test
Words like should, ought, fair and too much usually signal a normative statement, but check the meaning. “A 4-dollar cone is a rip-off” is normative, with no clue word. “Kai’s sales should rise on Sunday” is positive: here should means “probably will”, which Sunday’s sales can check.
Practice · 1 / 6
A classmate says: “Economic models leave out most of real life, so they’re useless.” Which reply is best?
Next
Like what you read?
Sign up free to save your progress, get spaced reviews tuned to you, and work through the whole course with interactive graphs and adaptive practice.