Setting Credit Limits - No Easy Answer
The financial controller of a computer hardware company a tough, competitive business with low margins and customers (computer retailers) that have been dropping like flies in the last year or so - asked me the other day how to set credit limits. The question came at the end of a hard day's seminar. I muttered something about "no easy answer" and "more of an art than a science". Feeling a little guilty at the inadequacy of that answer I thought I'd write something more comprehensive.
What are credit limits? They are a powerful risk management tool which allows you to limit the exposure - the amount that you might lose - on any customer. If you give XYZ Ltd a $100,000 credit limit then that should be the maximum you lose when they go belly-up. If you don't have a credit limit then some enthusiastic salesperson will go and sell them $200,000 worth of product the day before they go under. Initial credit limits are harder to set than existing limits are to review. New customers should generally expect to have to build up your confidence. As time goes on and a satisfactory payment history is developed, the credit limit can be increased and the risk reduction measures reduced.
Here's what I think is the best approach to setting credit limits, set out as a three step process.
- Sit down with the salesperson, look at past sales (if any) and project future sales over the next year.
- What will this mean in terms of maximum exposure? (Obviously this must take into account any seasonality in the industry, as for example the build-up of stock before the Christmas rush for many retailers.)
- Does the loss that you would make if the customer didn't pay after factoring in the probability of that loss exceed the profit that you expect to make. The statistical formula for calculating this is S P(n).n. To put it into English, the question is whether the sum of the probabilities of the outcome, times the outcome, reaches a result which is acceptable (i.e. makes you a satisfactory profit).
Let's look at a very simple example:
You (the creditor) expect to continue to sell $600,000 worth of product to an existing customer at a constant rate of $50,000 per month. The gross profit on this would be $60,000. The maximum exposure would be $100,000 (two months' sales). You want to work out whether to permit this credit limit.
Only 2% of your customers left you with bad debts last year but you think that the chance of this customer going bad over the next year is as high as 10%. To simplify things, we'll assume there are only two different outcomes. One is that the customer survives the year and you make $60,000. The probability of this, we assume, is 90%. The other is that the customer goes bust in the first month and you lose $100,000. We decide that the probability of this worst case scenario is 10%.
(Probability 1 x Outcome 1 [$60,000 profit]) + (Probability 2 x Outcome 2 [$100,000 loss]) = Expected Profit
( 90%x $60,000) + ( 10% x -$100,000) = Expected Profit
$54,000 - $10,000 = $44,000
The question of whether the predicted profit of $44,000 is acceptable is a business decision. In practice I'm sure most businesses would say that it is. (This is separate from the question of whether, from a cashflow perspective, the business could survive the loss of $100,000 in expected payments. If it couldn't survive, there are risk reduction techniques such as credit insurance which might be investigated.)
Clearly you could complicate this simplified approach in a variety of ways. You could assume different outcomes for each month rather than only two possible outcomes, and assume different probabilities of bad debt for different months. Many creditors will have some seasonality to consider and the possibility of mitigating your loss, say through recovery of product under your Romalpa clause, might also be taken into account. However, clearly, the really hard bit is working out the risk of the customer going bad and that is an art that would be the subject of a very much longer discussion than this column allows.
For those whose credit managers are daunted by this approach, carry out the first two steps then look at the alternatives below.
In the United States they have more practice at setting credit limits. They have some "easy answers" but they are not particularly good ones. In essence the Americans have two types of common guidelines. The simplest is to have some industry or company rule of thumb which relates the credit limit to the equity of the customer. For example, you might give a standard credit limit of 10% of the customer's equity. (Obviously, most creditors will then move from there, based on their experience of dealing with the customer.)
The problem with that is that there are companies with good cashflow but low equity and failing businesses with very high equity. The "percentage of equity" approach doesn't help much with these. And what if the customer wants to spend more than this approach would allow? Do you turn away the business? Ultimately the credit limit is a business decision. If the business's decision-makers want to take the risk, it's their prerogative to do so. The fact that the business has less equity than you would like is definitely not (on its own) a sure-fire pointer to the fact that the customer will default on your account.
A better and more sophisticated version of this is to set basic guidelines for different classes of customers based on equity or any other factors you choose. For example, you might have a small business limit, and a large corporate limit. You might classify your customers as A, B or C risks and have different standard credit limits for each class of risk. Or you might have standard limits for different industries. You then deal with exceptions on a case by case basis.
The other thing American credit managers do is look at what their competitors are doing. Industry credit groups are much stronger over there they do exist in New Zealand in some industries but in the US every industry has a credit group and credit managers from competing companies exchange information more readily. The credit manager from, say, IBM, can probably find out that Compaq has given XYZ Corporation a $500,000 credit limit. He then feels some justification in at least matching that limit.
There is an element of herd mentality in this. (It may just mean that everyone loses the same amount.) It fails to take into account the fact that some suppliers will be selling more to the customer. If sales of your product make up, or are forecast to make up, 10% of an existing customer's cost of goods sold (as determined from the COGS in their financial reports and your sales figures) while your competitor's product only makes up 5% of cost of goods sold (as determined by industry analysis or your sales manager's best guess) then other things being equal, your credit limit should be twice that of your competitor.
To summarise credit limits then: there is some science that can be partially applied, but my off-the-cuff answer was right; it's not particularly easy.
Chartered Accountants Journal of New Zealand