Economix Blog: Forecasting Unemployment

The unemployment rate for September surprised economists by dropping to 7.8 percent – the lowest rate since early 2009 – from 8.1 percent. But where will it go in the next month?

According to two economists with an innovative and impressively accurate new forecasting model, the answer is: nowhere. Expect the unemployment rate to stick pretty much where it is for months. And they say that with unusual confidence for forecasters, too.

Source: Brookings Institution

The economists, Regis Barnichon of the Barcelona Graduate School of Economics and Christopher J. Nekarda of the Federal Reserve, presented their new model in the Brookings Papers on Economic Activity series last month.

It is about 30 percent more accurate than other models on a three-month time horizon and performs especially well in the tricky times when the economy is in a recession, or turning around.

Here’s how it is different. Many unemployment-forecasting models rely on estimates of economic growth, or other macroeconomic data. But the Barnichon-Nekarda model looks at labor force flows, like the number of workers becoming unemployed, leaving the labor force and entering the labor force.

Let’s give an (extremely simplistic) example to show why the measure might be more precise.

Let’s say, for instance, that the pace of economic growth declined. A traditional model might predict that the unemployment rate would rise, with employers picking up fewer workers on the expectation of lower sales.

But let’s say that as growth declined more young people elected to stay in school and more workers in their 60s decided to retire. The labor force would shrink, meaning the unemployment rate might stay steady.

The Barnichon-Nekarda model might capture that near-term churn in the labor market more accurately than the traditional model – letting it more accurately forecast the unemployment rate.

So what does the Barnichon-Nekarda model tell us about September’s surprising drop? Don’t be so surprised, in essence.

“Unemployed individuals found jobs at a higher rate than in the previous month,” Mr. Barnichon writes in a blog post over at Brookings. “The rate at which workers lose jobs and enter unemployment also displayed a large unexpected drop. Consequently, the large unexpected drop in unemployment owes to very strong hiring and few layoffs.”

And what does the forecast see going forward?

Well, the economists do not foresee a correction to a higher unemployment rate in the coming months, given the pace of workers getting hired and fired. But they do not expect the rate to keep dropping – not at the pace of the last month, and really not at all. Much as the unemployment rate stubbornly stuck between 8.1 and 8.3 percent for most of this year, the economists see it sticking around 7.7 to 7.8 percent into the spring.

Article source: http://economix.blogs.nytimes.com/2012/10/05/forecasting-unemployment/?partner=rss&emc=rss