One of the problems with financial reporting is that even though it’s called ‘news’ the reports are often quite old. For example, the first estimate of GDP growth isn’t released until a month after the end of the period. So for example, the second quarter starts in April, ends in June, but doesn’t get reported until early July — and that July report isn’t actually a report of GDP, it’s the first of two estimates of what the actual report will be (which will be released almost two months later). This year it will be September 30th before we get the actual GDP number.
So roughly a month after the quarter ends, we get a first estimate of what happened during it. Roughly a month after that we get a second estimate. Roughly a month after that, we get the actual number. We just got the second estimate last week. This means that there is a lot of staleness in this number. Some of the economic activity being reported is from the beginning of April, almost five months ago.
It’s bad, but it’s an amalgam of data that is from five to two months out of date. How bad? Very bad, worst ever recorded.
That’s almost -33%. Ouch.
But what happens if we turn up the resolution on our microscope a little bit? It’s kind of low resolution to reduce three months of economic activity down to only one number. The word ‘analyze’ literally means to break or cut apart, so below we break the quarters down into months using a coincident indicator (meaning one which more closely coincides with the time period it is measuring) which is closer to real-time. In this case we use surveys of purchasing managers describing current conditions in their industry.
(Source: BEA, Institute for Supply Management, Q2 2020)
What we see above are three bar charts on top of each other. The first is GDP percent change. It shows a decline in output for the first quarter and a massive decline in the second quarter. But the second bar chart is a survey of service sector managers and the third line is a survey of manufacturing managers. They are called ‘high frequency’ statistics because they come out more frequently than GDP (every month rather than every third month). By overlaying these three sets of statistics we see that almost all of the contraction occurred in April and a bit of it in May. After that, conditions returned to normal. This means that the GDP data which just came out, which shows a terrible contraction, is largely about things which were happening four months ago. Interesting to know, and useful for historical analysis, but not really news in the sense that it’s not really new.