Avalanche Or Snowball: Which Debt Method Actually Wins

I carried a 22 per cent credit card balance for eight months while also putting money into a brokerage account. That was math I got backwards.

On a $4,000 balance, 22.9 per cent is $916 a year going out. A good year in the brokerage account might have returned 7 per cent, which on the same $4,000 is $280 coming in. I was paying $636 a year for the feeling of being an investor.

That is a different mistake from the one this piece is about, but it is the same family of mistake. I had five things to pay and I was not looking at them side by side.

The five debts

These are close to a real set. One store card opened at a till for a discount, one main card, a second card from a balance transfer that did not go to plan, a car loan, and a personal loan.

Look at the interest column rather than the balance column. The main card is quietly costing $128 a month, which is more than the car loan at nearly twice the balance, because 23.99 per cent on $6,400 beats 7.40 per cent on $11,200.

That one column is the whole argument, and almost nobody writes it down.

One honest caveat about the model

Everything below holds the minimums flat at $675 a month for the full run. Real card minimums shrink as the balance falls, because they are usually calculated as a small percentage of what you owe plus that month's interest.

Holding them flat makes both methods finish a little faster than they would in life, and it keeps the comparison between them clean, because the same simplification applies to both. The $339 gap below is sound. The 27 months is optimistic by a month or so.

The fix in your own life is to keep the total payment fixed at $1,000 no matter what the minimums do. That is what the model is really assuming, and it is also the right way to run it.

The budget, and what happens without one

The minimums come to $675 a month. Add $325 found by cancelling things and eating at home more, and the budget is $1,000 a month against $22,730.

Paying only the minimums on all five, forever, clears the balances in 82 months and costs $10,148 in interest. That is the baseline everything else gets measured against.

Avalanche: highest rate first

Pay the minimum on all five. Every spare dollar goes to whichever debt carries the highest rate. When that one hits zero, its minimum joins the pile and the whole amount moves to the next highest rate.

On these five debts, avalanche takes 27 months and costs $3,431 in interest.

What it feels like is the part nobody mentions. The store card dies in month 3, and then nothing happens for fourteen months while the $6,400 main card grinds down. Month 9 looks exactly like month 8.

Snowball: smallest balance first

Same mechanic, different queue. Minimums everywhere, spare dollars to the smallest remaining balance regardless of rate.

On these five debts, snowball takes 27 months and costs $3,770 in interest.

It feels completely unlike the other one. Store card gone in month 3, personal loan gone in month 6, second card gone in month 12. Three debts closed inside the first year, and three fewer statements arriving.

The gap

Three hundred and thirty nine dollars over 27 months is $12.54 a month, because 339 divided by 27 is 12.5. That is the entire prize in the argument people have about this.

The mechanism is easy to see once you have the interest column. Snowball leaves the main card sitting at $6,400 for five months longer than avalanche does, and that card accrues about $128 a month. Five months of roughly $128 is roughly $640 of extra interest, partly offset by clearing the smaller balances sooner, and the net is $339.

Which means the gap is not a fixed property of the two methods. It scales with how much money snowball leaves parked in a high rate debt and for how long. If the $6,400 card were $16,000 instead, the same five extra months would cost three times as much and the argument would deserve the energy people give it.

So which one

If you have never abandoned a repayment plan, run avalanche. It is arithmetically better by definition and you will finish it.

If you have abandoned one before, run snowball, and stop apologising for it. Three closed accounts in the first year against one is worth more than $12.54 a month to a person who needs evidence that the plan works.

A plan you follow at 90 per cent efficiency beats a plan you quit in month nine at 100 per cent efficiency.

The variable that actually mattered

Look at the two baselines again. Minimums only: 82 months and $10,148 of interest. Either method with an extra $325 a month: 27 months and roughly $3,500 of interest.

The extra $325 bought 55 fewer months and $6,717 less interest, because 10,148 minus 3,431 is 6,717 and 82 minus 27 is 55. It cost $8,775 in extra payments over 27 months, all of which went to balances rather than to a lender.

The order is worth $339. The extra money is worth $6,717. People argue about the first one because it is a debate and finding another $325 a month is just unpleasant.

Where the $325 came from

Nobody finds $325 a month in one decision. It came out of five places, and four of them I did not miss.

• Two streaming services and a music tier nobody in the house had opened in months, $41.
• Lunch bought out most days instead of twice a week, which at about $14 a day for twelve fewer days a month is $168.
• A phone plan I had not repriced in four years, $28.
• A gym membership I used eleven times in a year at $47 a month, which works out at $51 a visit.
• $47 of rounding and small things, which is where the last of it always comes from.

The lunch line is the only one that cost me anything in quality of life, and it stopped registering by about week three.

The car loan question

The car loan at 7.40 per cent sits last in both orders and stays there for 27 months, which bothers people because it is the largest balance on the list and it is costing $69 a month.

Leave it there. A fixed instalment loan at 7.40 per cent with a known end date behaves nothing like a revolving card balance at 23.99 per cent, and paying it early does nothing for your credit utilisation either, because a loan is not a card and does not count toward that figure.

The one thing to check is whether the loan carries a prepayment penalty, meaning a fee for clearing it ahead of schedule. Most do not. Find out which kind you have before the question becomes urgent.

Two levers before you start

Call each card issuer and ask for a lower APR, meaning annual percentage rate, the yearly cost of borrowing on that account. This is a five minute phone call with a real hit rate for anybody who has been paying on time, and the worst outcome is a no.

A balance transfer moves debt from one card to another at a promotional rate, usually for a fixed number of months, in exchange for a transfer fee of 3 to 5 per cent. On $6,400 a 3 per cent fee is $192. If the promotional window clears the balance, $192 to stop paying $128 a month is a good trade. If it does not, the leftover balance reverts to a rate that is often worse than where it started.

Do not transfer a balance you have no plan to finish inside the promotional window. That is how the second card in my table came to exist.

What to do with the cards once they hit zero

Leave them open. Closing a card you have just cleared removes its credit limit from your total, which pushes your utilisation up on the day you finally earned the right to have it low.

The store card is the exception worth thinking about, because a card designed to be used at one till at a discount is a card you will use at one till at a discount. Cut it up and keep the account open if you do not trust yourself with it in a wallet.

Before you pick anything

Open a note on your phone and write five lines, or however many you have. The name of each debt, the balance, and the rate. Then add a fourth column with balance times rate divided by 12, which is what each one is costing you this month.

Most people have never seen those four columns next to each other. The order you should attack in tends to become obvious within about thirty seconds of writing them down.

Mine did.