How Large Should a Checking Account Buffer Be? Margin & Month Questions about how much money should sit in a checking account often produce four confident answers, none tied to a particular household's timing. One month of expenses. Five hundred dollars. One paycheck. Whatever feels comfortable. Any one of those could fit one household and fail another. Here is one author-created stress test for estimating a possible checking cushion; it is not a CFPB formula or a universally correct balance. It uses two measured risks plus one deliberately chosen margin: how far bills have moved, how far income has moved, and an explicit allowance for what the history may have missed. This is Margin and Month. Today the fictional Bellamy household estimates a possible cushion, breaks that estimate into examples, and records the assumptions so the calculation can be reviewed later. The Bellamy household is fictional, as are every amount and date in this episode. This is general financial education, not advice about your accounts or your bank. First, separate two things that get called the same word, because mixing them produces a number that fails at both jobs. A checking cushion is intended to absorb ordinary variation: a higher bill, a delayed deposit, or a forgotten draft. In the Bellamy illustration it remains in checking because that is where the modeled variation occurs. An emergency fund is intended for larger losses or shocks: a job ends, hours are cut, or a car needs a transmission. Some people keep emergency savings separate to reduce accidental spending; an appropriate location depends on access needs, fees, insurance coverage, account features, and personal circumstances. Keeping both in one pool can make emergency money easier to spend unintentionally unless the two purposes are tracked separately. Today is entirely about the first one. The buffer. The Consumer Financial Protection Bureau defines an overdraft as a transaction the institution pays when the account does not contain enough money to cover it. The Bureau notes that checks, ATM transactions, debit-card purchases, automatic bill payments, and electronic or in-person withdrawals can overdraw an account. Some programs may charge a fee in addition to requiring repayment; other coverage can involve a linked account, credit card, or line of credit. It also says banks and credit unions have different fees and recommends obtaining the account-specific terms for overdraft protection, linked accounts, and lines of credit. We are not going to quote a fee figure. A household can use its institution's current disclosures, published fee schedule, or a confirmed answer from the institution. That account-specific information belongs beside any estimate of a possible cushion. Another account-specific question is whether the account has a linked-account transfer or an overdraft line of credit attached, and on what terms. The Bureau lists both as ways an overdraft can be covered. They may carry transfer fees, interest, or other costs, so the current terms determine what a shortfall costs in practice. Now the Bellamys build the illustrative estimate. It uses two measured risks plus one deliberately chosen margin. The first measured risk is variation in bill amounts. The Bellamys mark each variable line on their bill calendar — electricity, water, gas, and other usage-based charges — then use the median of every billed amount observed within the twelve-month period as the defined typical amount and record the highest bill in that period. A monthly bill may supply twelve observations; a quarterly bill ordinarily supplies four. For the Bellamys, electricity typically runs one hundred eighteen dollars and peaked at two hundred eleven in an August. That is a gap of ninety-three dollars. Their quarterly water bill typically runs sixty-one and peaked at ninety-six, a gap of thirty-five. The gaps total one hundred twenty-eight dollars. This is a deliberately conservative stress scenario because it assumes the separate highs could occur together; it is not a prediction or a measured aggregate monthly overage. A household could instead calculate the largest observed total monthly overage across all variable bills. Twelve months includes more seasonal variation than a six-month window confined to mild weather. A note on how the Bellamys read the twelve-month high: they distinguish ordinary variation from anomalies such as a leak, meter error, or one-off correction. An anomaly is documented, then considered separately when deciding whether it belongs in this cash-flow estimate or in an emergency plan. The point is to document both the number and the reason instead of silently discarding an inconvenient month. The second input is timing risk, and it is measured in days before it is measured in dollars. The Bellamys examine how late a deposit has plausibly become available. The Bureau's checking-fee guidance recommends confirming when deposited funds become available before treating them as spendable. So put a number of days on it. For the Bellamys, pay lands on alternate Fridays, and between a Monday holiday and a deposit hold they judge the realistic worst case at four days. Then the Bellamys total all scheduled payments inside that four-day window. Their rent, eleven hundred fifty dollars, drafts on the first. If an expected deposit slid four days, the rent money would need to have been reserved from earlier available income. So their modeled timing exposure is eleven hundred fifty dollars: the total of all scheduled outflows inside the window. This component is money reserved from earlier income for bills due before the next available deposit, not an additional uncommitted cushion. In the Bellamy illustration, this timing component is larger than the amount-variation component: a rent payment is modeled against an unavailable paycheck. The third input is a deliberately chosen and documented margin because historical measurements do not cover every future event. The Bellamys choose ten percent of the first two inputs, rounded up to a hundred thirty dollars, and label that choice rather than presenting it as a sourced rule. That gives an illustrative pre-rent required balance of one thousand four hundred eight dollars: one hundred twenty-eight dollars from the amount stress scenario, eleven hundred fifty dollars already committed to rent, and one hundred thirty dollars of margin. For a household whose largest bill lands before its next available deposit, the timing component can converge on the size of that bill. Because the Bellamy figure is built from stated assumptions, each input can be revisited when circumstances change. If avoiding a maintenance fee is part of the goal, the Bellamys would check whether their projected low balance remains above the account's required minimum. The Bureau lists going under a minimum balance requirement as one possible way to trigger fees. The one thousand four hundred eight dollar balance includes the rent money: eleven hundred fifty dollars is committed, while two hundred fifty-eight dollars is the amount-stress allowance plus the chosen margin. Savings earmarked elsewhere and available credit are excluded rather than counted as cushion. Consider two fictional households with identical income and bills but different timing for the largest payment. In the first, rent drafts after the paycheck has reliably become available, including the household's observed delay allowance. Their modeled timing exposure is close to zero, so the illustrative estimate contains mainly amount variation and margin. In the second, rent drafts three days before payday. The model includes the whole rent payment as timing exposure, making its illustrative estimate more than a thousand dollars higher. Same income and bills, but different dates produce illustrative estimates that differ by over a thousand dollars under this author-created model. This is why a universal number is unsatisfying. For the second household, one option is to ask whether changing a payment date is available and what it would cost. One option is to investigate timing changes first, then compare their availability and cost with holding a larger cushion. If fourteen hundred dollars is not immediately reachable, the Bellamys could choose interim milestones. The Bellamys choose three illustrative milestones; another household might choose differently. The first milestone is the amount stress scenario alone — one hundred twenty-eight dollars for the Bellamys. It reduces the modeled risk from an ordinary bill being higher than planned. The second is the peak of the single largest variable bill, so one bad August does not cascade. The third is the full one-thousand-four-hundred-eight-dollar illustrative estimate, which covers the scenarios modeled here without guaranteeing that an overdraft cannot occur. While below their illustrative estimate, the Bellamys compare available controls. The Bureau's checking-fee tool suggests asking whether a low-balance alert is offered and what it costs. A possible due-date change can also be evaluated for availability, fees, and effect on timing. If the landlord permits a no-cost date change, moving the draft could reduce the modeled timing exposure. A derived number goes stale, which is why we wrote down where it came from. The Bellamys review the inputs after a new variable-bill high, a pay-timing change, a move, a lease or premium change, or their scheduled annual date. They record that date beside the arithmetic. An optional exercise is to reproduce the Bellamy stress test: list variable bills, record the typical amount and twelve-month high for each, and add the gaps as a conservative scenario. The worksheet then records an observed deposit-delay allowance and reason, totals scheduled payments inside that window, and adds a labeled margin before dating the illustrative calculation. The account-specific companion information is the institution's current overdraft terms and whether a low-balance alert is available. There is no universal checking cushion. A number untethered from payment dates, bill variation, and account terms leaves out important information. The Bellamy illustration uses two measured risks plus one chosen margin: historical bill variation, observed deposit delay, and an explicit allowance. It is an author-created stress test, not a CFPB formula or a guarantee. Next time, costs that a selected month of statements may miss: the fictional Navarro household's year-ahead list of dated, sourced estimates.