[ad_1]
Michael Haddad
Text size
Uh Oh. The Dow lost 327 points as US stocks were defeated on Thursday. Global concerns shook the market after Chinese stocks were hammered overnight. Treasury Secretary Steven Mnuchin has withdrawn from a conference on investments in Saudi Arabia and the European Commission has challenged Italy's budget. Technology stocks like Netflix and
Facebook
was hit hard. Today After the bell, we …
- … marvel at the decline in the stock market;
- … wonder if a recession is imminent;
- … and look into Big Data to help answer that question.
What machines tell us about the business cycle
The market again fell sharply – and there was not a single problem to blame. the
S & P 500
ended down 1.4% to 2768.78 on Thursday, while the
Dow Jones Industrial Average
down 327.23 points, or 1.3%, to 25,379.45, and the
Nasdaq Composite
lost 2.1% to 7485.14.
In trying to determine whether this was a slight correction or the beginning of a bear market, investors had been trying to predict when a recession – the only thing known to kill a bull market – would occur. In this spirit, Aditya Bhave from Merrill Lynch, Bank of America has launched a machine learning algorithm in a monthly database of more than 100 economic variables to get an overview.
The computer has been asked to rank each month since the 1960s in three groups based on this vast set of macroeconomic data. The results were classified into two categories: economic growth, recessions and temporary expansions. And he found that the current cycle is really different from the others. Between the 1960s and the early 1980s, most expansions were in the category of booms. But from the mid-1980s, soft-patch type extensions became more common.
And this is especially true for the current cycle in which we find ourselves. As of July 2009, the economy was essentially stuck in the category of soft-patches, with only the month of last February being in the booming cluster. It's longer than all economic cycles of the post-war period except two. "Although [the current cycle] This is the second longest history of the post-war period, it has been essentially a very long period of time, with peak conditions of the business cycle so far elusive, "Bhave writes.
So where are we now? Over the last few quarters, the data is actually moving away from the recession cluster and getting closer to the expansion cluster, Bhave wrote. "In other words, the economy is far from recession."
Can this algorithm tell us when will the next recession take place? Bhave thinks so. Although the economic data used in the algorithms are three months behind schedule, it was nevertheless possible to point out the last three recessions – in 1990, 2001 and 2008 – more than three months in advance. In contrast, the National Bureau of Economic Research has qualified the last three recessions more than three quarters after their start, according to Bhave.
Will we trust the machine?
Write to Evie Liu at [email protected]
Source link
