Deflation and Disinflation: Falling Prices Against Slowing Inflation, and Why the Difference Matters

Deflation is the general price level falling, so a rupee buys more this year than last. Disinflation is inflation slowing while prices are still going up, just more slowly than before. The two words sound like variants of one another and describe opposite situations for anyone carrying a debt. The difference is worth insisting on.

Left on its opening figures the instrument works one year of the Republic of Sankhya. An index of 106.700 a year ago and 108.834 now, a rate of 2.00 per cent printed for the latest period against 6.70 per cent for the period before, a borrowing of Rs 10,00,000/- and an income that went from Rs 5,33,500/- to Rs 5,44,170/-. Since 106.700 multiplied by 1.02 is 108.834, the build reconciles exactly. The rate came down 4.70 percentage points and the level went up 2.00 per cent. The panel names that year disinflation. The Rs 10,00,000/- owed is worth Rs 9,37,207/- in base period rupees a year ago and Rs 9,18,831/- now, so the real burden eased by Rs 18,376/-, and a basket that cost Rs 4,000/- costs Rs 4,080/-. The rest of this guide works with those figures.

The confusion comes from two different quantities reported in the same sentence. One is the price levelA single number standing in for what a whole basket of goods and services costs, usually written as an index against a chosen starting year., the number for how much things cost. The other is the rate, a measure of how fast that level is changing. A rate can fall while the level it describes keeps climbing, the way a train slowing down is still moving away from the station. Almost every reader who gets this wrong has read a fall in the rate as a fall in the level.

What is deflation?

Deflation is a sustained fall in the general price level. Not a fall in the price of onions, not a fall in the price of televisions, but a fall in the broad measure that stands for what a basket of goods and services costs. If the indexA number that tracks how a whole group of prices has moved since a chosen base year, which is set to 100 so later readings are easy to compare. that measures the Republic of Sankhya reads 108.83 one year and 107.75 the next, Sankhya has had deflation. A rupee at the end of that year buys about 1.01 per cent more goods than it did at the start.

Notice what moved in that sentence: the level. Deflation means the level of prices went down, and nothing about a rate has to be mentioned at all to say it. A household in Sankhya paying Rs 4,000/- a month for its groceries and paying Rs 3,960/- for the same basket a year later has felt deflation, whatever any published rate happened to say that year.

Individual prices fall all the time without any of this being involved. Tomatoes crash after a good harvest. A four year old phone model gets cheaper every quarter. Because deflation is about the general level, none of that counts. The general level is dominated by everything else in the basket while those individual items move.

What is disinflation, and how is it different from deflation?

Disinflation is inflationA broad and lasting climb in what a basket of goods and services costs, so the same money buys less as the years pass. The causes of a climb like that are a separate subject. slowing down. Prices are still rising. Prices are simply rising less quickly than they were rising before. In the Republic of Sankhya, a headline rate of 6.70 per cent in one year followed by 2.00 per cent in the next is disinflation. Nothing in that pair of numbers is a fall in prices. Both are positive. Both describe a level that went up.

Now do the arithmetic that readers skip. Skipping it is where the whole error lives. Start the index at 100.00. A year at 6.70 per cent takes it to 106.70. The next year at 2.00 per cent multiplies 106.70 by 1.02 and takes it to 108.83. The rate fell by 4.70 percentage points and the price level ended 2.00 per cent higher than it started, so disinflation is a slowdown in a rise and never a fall. Somebody who heard that inflation had come down and expected their grocery bill to come down with it has been handed a bill that is 2.00 per cent larger.

The everyday version is a scooter going up a hill. In the first year it climbs fast. In the second the rider eases off and it climbs slowly. The rider has slowed and the scooter is still further up the hill. Slowing down is a statement about speed; being further up the hill is a statement about position. Disinflation is the first. Deflation is going back down.

Where do the two sit on a single axis of price movement?

Put everything a price level can do on one line and the confusion resolves itself: rising fast, rising slowly, not moving, falling. The four positions cover every case, and each describes where the level is going rather than how it got there.

Sort them and the words attach cleanly. Rising fast is inflation. Rising slowly is inflation. Not moving is the nil boundary, where neither word applies. Falling is deflation. Only two of those four positions are inflation and only the last is deflation. Disinflation is a movement from one position toward another, and a movement is not a position. Disinflation also cannot be read off a single year. Comparing one rate against the rate before it takes two years.

Disinflation is a move leftward along that axis, from rising fast to rising slowly, and it can keep going until it reaches the nil boundary. The move can carry on past the nil boundary, and at the moment the rate crosses nil the economy has stopped disinflating and started deflating. The two are neighbours on a path rather than synonyms. Neighbours on a path are easy to muddle.

Can an economy have disinflation and rising prices at the same time?

Yes, and it is the ordinary case rather than a curiosity. Disinflation is rising prices, by definition. A rate that is falling but still positive describes a level that is still going up. There is no version of disinflation in which prices are not rising. The question sounds like a paradox only because the word disinflation has the word deflation hiding inside it, and the ear supplies a meaning the arithmetic never had.

An economy in disinflation is an economy whose prices are rising, and the only thing that has slowed is the speed of the rise. Sankhya at 6.70 per cent and then 2.00 per cent is in disinflation for the whole of that second year while its price level sets a new high in every single month of it. The two statements do not compete: one is about the rate, one is about the level.

What does the Sankhya price level actually do across four years?

Definitions settle nothing until the level is walked alongside the rate, so here is the walk. Start the index at 100.00 in year 0. Apply the headline rate of 6.70 per cent and year 1 closes at 106.70. Apply 2.00 per cent and year 2 closes at 108.83. Apply minus 1.00 per cent and year 3 closes at 107.75.

The rate column falls in every year: 6.70, then 2.00, then minus 1.00. The level column rises, rises, and only then falls. The rate fell three years running and the level went up in two of those three. The mismatch between those two columns is the entire distinction between disinflation and deflation.

One arithmetic warning about that table, of the sort that quietly corrupts a spreadsheet. The levels are computed by compoundingApplying each year's growth to the result of the previous year rather than to the original starting figure, so the effects multiply instead of adding. the full precision figures and rounded only for display. Year 2 is really 108.834 and year 3 is really 107.74566, displayed as 107.75. Applying minus 1.00 per cent to the displayed 108.83 instead gives 107.74, adrift on the index and further adrift on every rupee figure built from it. Rounding belongs at the end, never in the middle.

Why is falling prices treated as a problem rather than a relief?

Because two mechanisms run underneath it, and both work against the household that is supposed to be enjoying the cheaper prices. Neither mechanism is obvious from the price tag. Deflation feels like good news for exactly as long as it takes to think it through.

The first mechanism is deferred spendingPutting off a purchase already planned, usually because the same thing is expected to be cheaper later.. If a household expects a washing machine to cost less in six months, waiting is rewarded. Waiting is rewarded for the neighbours too. So the shop sells fewer washing machines, cuts its price to move stock, and confirms exactly the expectation that made everyone wait. Nobody is behaving strangely; everybody is behaving sensibly given what they expect. Falling prices reward waiting, waiting reduces buying, and less buying pushes prices down further.

The second mechanism is more mechanical and does not depend on anybody expecting anything. A debt is fixed in rupees, so when the rupee gets stronger the debt gets heavier without a single term of the agreement changing. A household with a loan of Rs 10,00,000/- and an income that moves with the price level finds that its income shrinks in a deflation while the Rs 10,00,000/- does not shrink at all.

What happens to a debt of Rs 10,00,000/- when prices fall?

The debt gets heavier, and nothing about it changes to make it so. Take a Sankhya household carrying a nominal obligationAn amount owed that is written in a fixed number of rupees, so the figure on the agreement stays the same no matter what happens to what a rupee buys. of Rs 10,00,000/-. Assume an income of Rs 5,00,000/- a year in year 0 that moves exactly with the price index: if the index rises 6.70 per cent the income rises 6.70 per cent, and if it falls 1.00 per cent the income falls 1.00 per cent. The assumption that income tracks the index exactly is a simplification. Real incomes track the price level loosely, late and unevenly.

Now watch the real burdenWhat an amount owed is worth in terms of goods rather than in rupees, found by dividing the rupee figure by the price index so that different years can be compared. move. Dividing the Rs 10,00,000/- by the index against the year 0 base restates the debt in year 0 rupees. In year 0 that is Rs 10,00,000/-. By the end of year 1 the index is 106.70 and the debt is worth Rs 9,37,207/-. By the end of year 2 the index is 108.83 and it is worth Rs 9,18,831/-. Then deflation arrives, the index drops to 107.75, and the debt climbs back to Rs 9,28,112/-. The household repaid nothing and borrowed nothing, and its debt got Rs 9,281/- heavier in real terms the moment the price level fell.

The fourth column is the one to stare at: it never moves. Everything else in the table shifts around a constant Rs 10,00,000/-, and the whole household experience of that debt getting lighter and then heavier comes from those other columns. Across years 0 to 2 the real burden fell by Rs 81,169/-. In year 3 alone it rose by Rs 9,281/-.

One honest caution about checking that table twice

Treating the last two columns as two independent confirmations of the same finding is tempting. The two columns are not independent. The debt in year 0 rupees is Rs 10,00,000/- divided by the index. The debt as a multiple of income is Rs 10,00,000/- divided by an income that is itself Rs 5,00,000/- multiplied by that same index. One is the other divided by a constant Rs 5,00,000/-. Two rearrangements of one expression will always agree, and their agreement confirms nothing, so the only thing worth checking here is the index itself. A real second check would have to come from somewhere else, such as a different measure of the price level.

Does inflation do the same thing in reverse?

Inflation does the same in reverse, and seeing the mechanism run both ways stops the deflation story sounding like a complaint about one condition. Under inflation, an income that moves with the price level grows while a debt fixed in rupees stands still, so the borrower repays with rupees worth less than the rupees borrowed. Value has moved from the lender to the borrower. Under deflation the identical mechanism runs the other way. Deflation moves value to lenders and inflation moves value to borrowers. Neither direction is fair or unfair. Both are a fixed rupee obligation behaving as it must when the rupee itself changes.

The Sankhya table shows both directions in one place, and three years are the fewest that can show both. Across years 1 and 2 the real burden fell by Rs 81,169/-, and that Rs 81,169/- did not evaporate: it is value the lender no longer gets in real terms. In year 3 the direction reverses and Rs 9,281/- goes back the other way. Nobody negotiated either transfer, and the agreement said Rs 10,00,000/- throughout.

Lenders are not helpless in the face of this. A lender who expects the rupee to lose value asks for a higher rate at the outset, compensation for the expected change built into the price of the loan. The higher rate works only to the extent the change was expected. The transfer that surprises people is the part nobody priced.

Why is reading the rate instead of the level a costly habit rather than a pedantic one?

Because the rate is a statement about the change in the change, and almost nothing a household or a lender needs to decide depends on that. Costs depend on the level. The buying power of a salary depends on the level. Whether a debt got heavier depends on the level. The rate matters as an input to the level rather than as a substitute for it, and treating it as a substitute produces conclusions that are exactly backwards.

The difference between a smaller rise and a fall is not a fine point of vocabulary: it is the difference between paying more and paying less, and every plan built on the wrong one is built on a sign error. Somebody budgeting a household on the belief that costs will come down, when costs are going up by 2.00 per cent, will be short. Somebody lending against the belief that a borrower income will keep growing with prices, when the price level has turned down, has mispriced the risk in the direction that hurts. The habit that prevents both: before saying anything about what things cost, find the level.

What does a lender actually watch when prices move?

Not the inflation rate on its own. The rate tells a lender very little about any particular borrower. A lender watches whether the borrower revenue moves with the price level while the borrowing does not. The gap between the two is where the risk sits in either direction. A trader whose selling prices move with the general level and whose loan is fixed in rupees is comfortable in inflation and exposed in deflation. A salaried borrower on a contract renegotiated every three years has an income fixed for a while too, and is exposed for most of that period whichever way the level went.

The same reading applies to a business. A credit assessor looking at a Sankhya trading firm with Rs 10,00,000/- of fixed rate borrowing asks how much of the firm revenue reprices with the general price level, and how quickly. A firm selling on annual contracts reprices slowly. A vegetable wholesaler reprices daily. Neither is better; they sit at different points on the same exposure, and a fall in the price level lands on the slow repricer hardest and lasts longest there.

An analyst asks the paired question about the other side of the balance sheet. Cash and fixed rupee receivables gain in real terms when the level falls and lose when it rises, the mirror image of the borrowing. The useful question is never whether prices are rising or falling but which of a firm claims and obligations are fixed in rupees and which move with the level. The framing survives both conditions. No rule of thumb attached to one of them does.

The Republic of Sankhya is invented.
Educational material. Not advice on any investment, tax, budget or market position.