Arkady
President
Over the years, I've had a lot of debates about inflation, and I find them interesting because they don't break down on the usual partisan lines. You can guess how the knee-jerk partisans are going to argue based on what's convenient for their party at the moment, but those arguments keep reversing themselves (the right-winger who wants to insist real inflation is much higher than the official rate, right now, would have made the opposite argument a few years earlier when higher inflation would have looked bad for Bush).
I sympathize with those who have to compile the official inflation rates, because there's no easy and fair way to do it. I wonder whether most people here have really thought through the challenges. Ultimately, what you're trying to record is how far a dollar will go for a consumer. So, what you do is you come up with a basket of goods, and figure out how much it costs to buy that basket. You want the basket to represent a fairly typical collection of purchases for an average consumer -- each type of purchase in relative proportion to the other purchases he makes, so that a move in one item reflect the overall inflation rate in proportion to how hard such a move would hit a typical consumer's bottom line.
That sounds pretty easy, and if what a typical consumer bought never changed, there'd be no difficulty or controversy involved in the inflation rate. E.g., if a typical consumer in 1960, 1970, 1980, 1990, 2000, 2010 bought the same average basket of goods each month (30 gallons of gas, two dozen eggs, four gallons of milk, two pair of underwear, twenty newspapers, etc.), then you could just price them out in each year and compare. But a typical consumer's purchases change over time, so how do you account for that?
Obviously, some of that happens due to technology changes. For example, a typical consumer today is far more likely to buy a color TV than a typical consumer in 1960 (when that would have been a luxury good). For a couple hundred dollars, today, I can buy you a smart phone with more storage and processing power than a $100,000 super computer would have gotten you in 1960. How do you account for that?
It's difficult to compare products across time periods even when the fundamental technologies aren't that different. The basics of the American automobile aren't all that different -- four wheels and an internal combustion engine. Still, to what do you compare a mid-market mid-sized sedan of 2012, when you want to price it out against an equivalent product in 1972? In theory, you could just compare it to a mid-market mid-sized sedan of 1972.... but chances are your 2012 model has vastly more going for it -- airbags, anti-lock braking, traction control, a far more fuel efficient engine, more horsepower, lighter materials, stability control, high-quality sound system, power windows, power steering, power locks, maybe even things like GPS.
People also make substitutions based on pricing, and those have to be accounted for. For example, shrimp farming has turned shrimp from a luxury items just a couple decades ago to something you get in dirt-cheap appetizers at low-end family restaurants. In 1998, a pound of shrimp cost $18.80. Today it costs $7.96. The predictable result is that people will eat more shrimp. That changes the proportional composition of shrimp within the "basket of goods" that a typical consumer purchases. On the flip side, the price of beef has sky-rocketed -- especially in the last few years. Yet people are eating less beef (partly due to price and partly due to health concerns that have made chicken and fish relatively more attractive). So, again, the typical "basket of goods" has changed. You could price out what it would cost for someone to buy what was typical in 1998, but it wouldn't mean that much if that's just not what a typical consumer is buying these days. Or you could price out what it would have cost in 1998 to buy what a typical consumer buys today, but again that isn't terribly meaningful if a typical consumer wouldn't have been buying that (e.g., lots of shrimp).
Anyway, I'm not posting this to defend how the government calculates inflation, and certainly not to argue that there aren't improvements to be made. I'm just saying that a lot of the criticisms I see are just superficial ones, driven by a desire to game the numbers one way or the other for political purposes. Most people haven't even begun to grapple with how hard it is to come up with a good inflation calculation. No matter how much a person wants to calculate the "real inflation rate," and no matter how good he is about dismissing his own partisan prejudices about it, there's no simple or uncontroversial "real" number to be had. As long as technology and consumption patterns keep changing, any calculation will be fraught with judgment calls and controversies.
I sympathize with those who have to compile the official inflation rates, because there's no easy and fair way to do it. I wonder whether most people here have really thought through the challenges. Ultimately, what you're trying to record is how far a dollar will go for a consumer. So, what you do is you come up with a basket of goods, and figure out how much it costs to buy that basket. You want the basket to represent a fairly typical collection of purchases for an average consumer -- each type of purchase in relative proportion to the other purchases he makes, so that a move in one item reflect the overall inflation rate in proportion to how hard such a move would hit a typical consumer's bottom line.
That sounds pretty easy, and if what a typical consumer bought never changed, there'd be no difficulty or controversy involved in the inflation rate. E.g., if a typical consumer in 1960, 1970, 1980, 1990, 2000, 2010 bought the same average basket of goods each month (30 gallons of gas, two dozen eggs, four gallons of milk, two pair of underwear, twenty newspapers, etc.), then you could just price them out in each year and compare. But a typical consumer's purchases change over time, so how do you account for that?
Obviously, some of that happens due to technology changes. For example, a typical consumer today is far more likely to buy a color TV than a typical consumer in 1960 (when that would have been a luxury good). For a couple hundred dollars, today, I can buy you a smart phone with more storage and processing power than a $100,000 super computer would have gotten you in 1960. How do you account for that?
It's difficult to compare products across time periods even when the fundamental technologies aren't that different. The basics of the American automobile aren't all that different -- four wheels and an internal combustion engine. Still, to what do you compare a mid-market mid-sized sedan of 2012, when you want to price it out against an equivalent product in 1972? In theory, you could just compare it to a mid-market mid-sized sedan of 1972.... but chances are your 2012 model has vastly more going for it -- airbags, anti-lock braking, traction control, a far more fuel efficient engine, more horsepower, lighter materials, stability control, high-quality sound system, power windows, power steering, power locks, maybe even things like GPS.
People also make substitutions based on pricing, and those have to be accounted for. For example, shrimp farming has turned shrimp from a luxury items just a couple decades ago to something you get in dirt-cheap appetizers at low-end family restaurants. In 1998, a pound of shrimp cost $18.80. Today it costs $7.96. The predictable result is that people will eat more shrimp. That changes the proportional composition of shrimp within the "basket of goods" that a typical consumer purchases. On the flip side, the price of beef has sky-rocketed -- especially in the last few years. Yet people are eating less beef (partly due to price and partly due to health concerns that have made chicken and fish relatively more attractive). So, again, the typical "basket of goods" has changed. You could price out what it would cost for someone to buy what was typical in 1998, but it wouldn't mean that much if that's just not what a typical consumer is buying these days. Or you could price out what it would have cost in 1998 to buy what a typical consumer buys today, but again that isn't terribly meaningful if a typical consumer wouldn't have been buying that (e.g., lots of shrimp).
Anyway, I'm not posting this to defend how the government calculates inflation, and certainly not to argue that there aren't improvements to be made. I'm just saying that a lot of the criticisms I see are just superficial ones, driven by a desire to game the numbers one way or the other for political purposes. Most people haven't even begun to grapple with how hard it is to come up with a good inflation calculation. No matter how much a person wants to calculate the "real inflation rate," and no matter how good he is about dismissing his own partisan prejudices about it, there's no simple or uncontroversial "real" number to be had. As long as technology and consumption patterns keep changing, any calculation will be fraught with judgment calls and controversies.
