If you have a budget--more specifically, if you track your spending--you have the data you need to track changes in your own personal cost of living. That's a lot more useful than the Bureau of Labor Statistics Consumer Price Index (although people on Social Security, or with a lot of money invested in inflation-indexed treasury bonds, care about the CPI too). It's not so easy to turn your data into a number, though. In fact, you'll face a lot of the same problems the government does.

The first thing to do is to think about what you're really trying to get at. If all you wanted to know was your cost of living, you could calculate that pretty easily. If you're tracking your spending, it's right there--the sum of all your spending. Even if you're not, you can come pretty close: Add up all your income, subtract out any money that went into savings or investments, and the difference is, more or less, what you spent.

Usually, though, you're trying to figure out something slightly more subtle. What you want to know is, am I getting ahead, or am I falling behind? Am I spending more because my standard of living has gone up, or am I spending more because the money is worth less?

In some cases, it's easy to decide. If you stayed in the same apartment, and your rent went up by 3%, that's probably almost all higher cost of living. (The exception would be if the apartment had been improved in some way, such as by swapping our your old refrigerator for a new energy-efficient model.)

In other cases, it can take thoughtful consideration and some careful calculation. Suppose your grocery bill is up by 7%. There's almost no way to know what that means in terms of your cost of living, unless you've kept accurate notes about things like how often you ate out and how often you fed other people. A 7% increase in your grocery bill could amount to a drop in your cost of living if you cooked one extra meal a week instead of eating out. It would be a huge drop if you went from feeding two people to feeding three with only a 7% increase in cost.

Basically, you're running into the same problems the government does. It calculates the CPI based on a standard "basket" of goods and services. That allows them to avoid the household size issues, but it has plenty of its own problems.

What goes in the basket?

Simply for practical reasons, the CPI price basket can't include everything anyone might buy. Government economists pick a basket with a reasonable number of items, trying to pick things that a good indicators of a whole class of items.

One classic issue is "substitution"--as items become more expensive, consumers buy less of that and more of something that's cheaper. If the basket were to simply follow what people buy, it would miss some of the rise in the cost of living. On the other hand, if something that used to be a standard item (let's say, hardwood molding, cut in a particular scallop pattern), is kept in the basket even after it becomes a speciality item, the price changes in that item reflect other things than changes in the cost of living.

The government has economists and bureaucrats trying to balance those trends. You have the advantage of being able to look at what you actually spend.

What's the price?

There are plenty of items that are always on sale, in some form. At many stores, for example, there is always one brand of soda on sale, always one brand of premium orange juice on sale, etc. Consumers who are price sensitive--and consumers who plan ahead and stock up--can always buy the one that's on sale. CPI data, though, is for specific items and is gathered at a specific time. Sale prices probably average out over time, but they may not, and they definitely introduce noise into the system.

There's also the issue of changes in where people shop. How much of the CPI should be based on the prices at WalMart, and how much should be based on the price at the little hardware store downtown?

Again, you don't have to worry about these issues, because you can calculate your personal CPI on what you actually spend, where you actually spend it.

Other issues

There are plenty of other issues that keep it from being a trivial exercise to calculate your own personal cost-of-living index.

Boundary cases--If some years you pay a bill in late December and other times you pay it in early January, you'll have some years that you don't pay it and other years that you pay it twice. That doesn't mean your cost of living is bouncing around.

Major purchases--If you bought an expensive flat-screen TV last year, but didn't buy another one this year, your "household goods" category might show a 50% drop, but it's hard to say just what that means in terms of your cost of living.

Gifts--What you pay for gifts that you give is probably part of your cost of living, but it's harder to decide how to handle gifts that you receive.

Worth doing

In the greater scheme of things, though, it almost doesn't matter how you decide to handle these sorts of special cases, because the results will be useful in any case.

If you go through the exercise, you'll definitely learn some things. If you just go with your overall impression, as opposed to actually cranking the numbers, it's easy to over-emphasize changes in the prices that are highly visible (gasoline, for example). Contrariwise, you're likely to under-emphasize price changes in things that seem to be bargains (even if it's still cheap, a 15% increase in the price of peanut butter is a 15% increase).

The Bureau of Labor Statistics released their consumer price index today. It was up 0.2% last month, up 3.5% from a year ago.

If any of you calculate the change in your cost of living, I'd be very interested to hear what numbers you come up with.