What it is:
How it works (Example):
The formula varies, but the most common one is:
Efficiency Ratio = Expenses* / Revenue
*not including interest expense
For example, if Bank XYZ's costs (excluding interest expense) totaled $5,000,000 and its revenues totaled $10,000,000, then using the formula above, we can calculate that Bank XYZ's efficiency ratio is $5,000,000 / $10,000,000 = 50%. This means that it costs Bank XYZ $0.50 to generate $1 of revenue.
As we said earlier, the formulas vary but the idea is to look at costs as a percentage of revenue. Costs include salaries, rent and other general and administrative expenses. Interest expenses are usually excluded because they are investing decisions, not operational decisions.
Revenue includes interest income and fee income, though some banks exclude their provision for loan losses from revenue or add their tax equivalent net interest income to revenue when calculating the efficiency ratio.
Why it Matters:
The efficiency ratio is a quick and easy measure of a bank's ability to turn resources into revenue. The lower the ratio, the better (50% is generally regarded as the maximum optimal ratio). An increase in the efficiency ratio indicates either increasing costs or decreasing revenues.
It is important to note that different business models can generate different efficiency ratios for banks with similar revenues. For instance, a heavy emphasis on customer service might lower a bank's efficiency ratio but improve its net profit. Banks that focus more on cost control will naturally have a higher efficiency ratio, but they may also have lower profit margins.
In addition, the more a bank generates in fees, the more it may concentrate on activities that carry high fixed costs (and thus create worse efficiency ratios). The degree to which a bank is able to leverage its fixed costs also affects its efficiency ratio; that is, the more scalable a bank is, the more efficient it can become. For these reasons, comparison of efficiency ratios is generally most meaningful among banks within the same model, and the definition of a "high" or "low" ratio should be made within this context.