Start from the record

Most questions about a market or an economy reach us wrapped in opinion. A headline says a sector is in trouble. A chart doing the rounds says a price can only go one way. A confident voice says the decision was obvious. Before we form a view of our own, we set all of that aside and go to the record: what was filed, published or said, by whom, and when.

What we mean by the record

The record is what the people closest to a question put down in public, in their own words and numbers. For a company, that means its annual and quarterly reports, its prospectuses and the statements it makes when it publishes them. For an economy, it means the statistics that national agencies publish on output, prices and jobs, and the decisions and minutes of central banks. For a market, it means prices and volumes as they were traded.

We prefer the first source to any later account of it. A news story about a filing is useful for knowing that the filing exists. It is not a substitute for reading it.

As it stood at the time

Much of the record changes after it is first published. Official statistics usually start as estimates and are revised as fuller information arrives. Companies restate past results. Indexes change their members, and the companies that failed drop out of today’s lists.

This matters whenever we ask what someone could have known on a given day. If we judge a past decision with figures that were published later, we credit people with knowledge they did not have, and we learn the wrong lesson. So when the date matters, we read the record as it stood on that date. Archives that keep every release of a statistic make this possible: the Federal Reserve Bank of St. Louis keeps one, called ALFRED, for a great many economic series.

Where the record misleads

Starting from the record is not the same as trusting it. Definitions change, so a series can break in ways its chart does not show. Seasonal adjustment smooths some patterns and can hide others. A company’s filing says what the company chose to say, within the rules. And a list of the survivors, whether funds, firms or coins, tells you little about the ones that did not survive.

So we read the record with its own caveats: the notes at the foot of the table, the definitions at the back of the report, the date of each revision.

How we use it

In practice the order is simple. We write the question down. We collect the record that bears on it and note the date of every item. Only then do we read what others have said about it. When the commentary and the record disagree, we say so, and we say which one we have relied on.

Starting from the record does not make us right. It makes our work possible to check, which is the first thing research owes its readers.