当我收拾自己的CD藏品（它们是荒废的青春正在腐烂的果实）时，我的麻烦来了。我指的不是把它们按字母顺序排列，而只是分拣一堆堆有划痕的银色盘，把它们一一放回CD盒。这一过程提醒我，有多少音乐没有听过，而原因只是需要进行这种“考古发掘”。因此，我开始考虑将这些音乐全拷贝到一些美妙的电子设备中，然后把CD扔掉。
经过一些研究后，现在我意识到，自己的选择多得令人眼花缭乱——媒体个人电脑、iPod基座、专用音乐服务器和无线“桥梁”。别让我开始列举它们的优缺点，不过有一点似乎非常清楚，即它们之中多数设备处理音乐的方式是我们之中多数人10年前难以想象的。人类的反应是困惑。当然，经济学家的反应是“我想知道这些小玩意中有多少计入了通货膨胀统计数据？”
当我还是孩子的时候，还没有CD。（CD今年刚满25岁，第一张CD是由Abba乐队发行的《The Visitors》。）最早的CD播放机贵得离谱，因此买的人并不多。只有当价格降下来之后，人们才开始接受CD格式。
对于编纂通胀数据的统计学家们来说，这加大了他们工作的难度。国家统计局(Office for National Statistics)每月会派遣调查人员去各个商店一次；调查员记下各类价格；国家统计局的某些人将所有价格加在一起，然后我们可以看到过去一个月的通胀幅度。多年来，CD播放机的价格不断下跌，质量却在不断提高，这应该对降低通胀有所贡献。
但是，当价格下跌而质量提高的幅度最大时，恰好是它们对通胀统计数据的影响很小甚至为零的时候。这是因为在CD播放机非常稀缺时，国家统计局没有搜集CD的价格。他们只是从今年开始搜集小型平板电视、数字收音机和数码照片处理等现代科技“奇迹”的价格。
当这些新产品出现时，通胀统计数据没有承认它们的存在，只有当它们价格下跌、质量提高到足以被人们广泛接受时，才纳入了通胀统计数据。当价格下跌幅度最大时，人们还没有开始大量购买，因此国家统计局尚未将这些新产品纳入其月度调查中。
目前，国家统计局已经搜集12万类价格（代表许多不同地区的650种消费产品和服务）。如果他们拥有更多的渠道，他们可能搜集更多价格，但这不会改变根本的困难。在计算过程中，新产品仍然只占很小的权重，直到它们成为大众产品为止。
这是表明通胀数据搜集工作中一个更普遍问题的特殊案例：只有我们大量消费法国布里乳酪时，它的价格变化才会体现在通胀数据中。而决定我们是否掏钱买这种乳酪的因素之一，当然是它的价格。价格变化影响我们购买的对象，而我们购买的对象又影响统计学家计算价格变化的方式。[（电视厨师）杰米•奥利弗(Jamie Oliver)可能也有影响。今年法国布里乳酪不在国家统计局的调查范围内；而橄榄油在。]
不要责怪国家统计局。要想计算过去50年的通胀程度，它必须指出多少部留声机相当于一部iPod。该问题无法解决，如果你对此有不同见解，我建议你直接写信给国家统计局。另一方面，如果你能告诉我该如何处置我的CD藏品，我愿意洗耳恭听。
Undercover Economist: Lost in music
My troubles began while I was tidying up my CD collection, the decaying fruit of a misspent youth. I don't mean alphabetising it, merely sorting through the piles of scratched silvery discs and putting them back into their cases. The process reminded me just how much music I don't listen to, simply because of the archaeological dig that would be required. And so I started to think of copying all this music into some wonderful electronic box, and chucking the CDs away.
After a bit of research, I now realise that I have a dizzying range of choices – media PCs, iPod docks, dedicated music servers and wireless “bridges”. Don't ask me to begin to enumerate the pros and cons, although it seems clear that most of them do things with music that it would have been hard for most of us to imagine 10 years ago. The human response is bewilderment. The economist's response, of course, is “I wonder how many of these gizmos are in the inflation statistics?”
When I was a boy, there were no CDs. (The CD is 25 years old this year; The Visitors, by Abba, was the first one produced.) The original CD players were ludicrously expensive, so not many people bought them. Only when the price came down did people embrace the CD format.
That makes things difficult for the bean-counters who compile the inflation statistics. The Office for National Statistics (ONS) sends surveyors out to the shops once a month; the surveyors take a note of the prices; someone at the ONS adds up all the prices, and we see how much inflation has taken place in the past month. CD players have been falling in price and improving in quality for years, which should show up as a contribution to lower inflation.
But it is precisely when the price falls and quality improvements are most dramatic that they make little or no impact on the inflation statistics. That is because when CD players were scarce, the ONS did not collect CD prices. Only this year did they start collecting the prices of modern marvels such as small flat-panel televisions, digital radios and digital-photo processing.
When such new goods appear, the inflation statistics do not acknowledge their existence, and only do so when their price has fallen and quality has risen enough to make them widely adopted. When prices are falling most steeply, people are not yet buying in bulk and the ONS is therefore not yet including the new goods in its monthly survey.
The ONS already collects 120,000 separate price quotations (representing 650 consumer goods and services in many different locations). If they had more resources, they could collect more, but it would not change the fundamental difficulty. New goods would still be given a tiny weight in the calculations until they became mass-market products.
It's a special case of a more general problem in collecting inflation data: a change in the price of brie shows up in the inflation figures only if we spend a lot of money on brie. But one of the things that determines whether we spend money on brie is of course – the price of brie. Price changes affect what we buy, what we buy affects how statisticians compute price changes. (Jamie Oliver probably has an effect, too. Brie is out of the ONS survey this year; olive oil is in.)
Don't blame the ONS. To calculate how much inflation has taken place over the past 50 years, it would have to show how many gramophones are equivalent to an iPod. The problem is unsolvable, and if you disagree, may I recommend that you write directly to the ONS. On the other hand, if you can tell me how to deal with my CD collection, I am listening.
