Forecasting that the Unemployment Rate will stay Constant is a Bad Idea

View fullsize

Jim Bullard, President of the St Louis Fed, has released a new, St Louis Fed model, for thinking about the way the Fed forecasts. According to the St Louis model, we should think about 'regimes'. There are three components to regimes. 1) Is the economy in a recession: YES or NO? 2) Is the short-term real interest rate HIGH or LOW? 3) Is productivity growth HIGH or LOW? 

Putting these pieces together, there are eight possible states. Recession can be YES or NO,  productivity can be HIGH or LOW and the natural real interest rate (Jim calls this R Dagger) can be HIGH or LOW. 

In Bullard's view the current regime is

Recession: NO,      Productivity growth: LOW,    R Dagger:   LOW

Using regime dependent forecasting, Jim thinks the best forecast of the US economy, moving forwards, is that productivity growth will stay low and the unemployment rate will stay where it is. That implies, according to Bullard, that the Fed should hold the interest rate at 63 basis points through 2018. 

I have one big problem with this forecasting framework. Take a look at the figure above which depicts US unemployment since 1950. Jim Bullard wants to talk about a 'regime dependent equilibrium'. I have no problem with that idea. But there is no period in the post-war period when the unemployment rate was even approximately constant. It was either increasing or it was decreasing.  If we stick with the regime dependent paradigm, I would replace, [Recession = Yes or NO], with, [Unemployment = INCREASING or DECREASING].

That may seem like a semantic change. But it makes a big difference to a regime dependent forecasting model because the unemployment rate cannot keep falling forever. That suggests that, the longer we are in the [Unemployment = DECREASING] state, the higher is the probability of a regime switch into [Unemployment = INCREASING]. That suggests to me, that the risk of another recession while productivity and the natural real interest rate are low is higher than Jim Bullard thinks. 

I'm glad that the St Louis Fed has moved to this new framework as I've argued for a long time that the existing paradigm is broken.

You can preorder my book, Prosperity for All, which says a lot more about these issues, HERE.

Is the US economy healthy?

I wrote this piece for  John Kiernan at Wallethub ... you can find how some other economists answered the same question here. 

The US economy is not dead, but it's not about to run a triathlon any time soon. We're in intensive care in a third rate hospital in the hands of a medical team struggling to come up with a convincing diagnosis. The good news is that the patient is resilient and the knowledge we gain from experimental cures  will  help us save future patients.

There are two problems. The first is the hangover from a binge lending episode that ended badly in 2008. That episode was caused by a bout of optimistic expansion in which half the world lent to the other half. Everyone knew the asset price boom would end. But the gains made during the upswing were too big to ignore. And when it came: the landing was hard. The borrowers are still paying down their accumulated debts.

The problem with the capital markets is not, as some would claim, that people are irrational. During the debt crisis, we were all acting in our own best interest. And we knew it. People are rational.  Markets are not. The world economy is recovering from a nasty bout of idiopathic rational exuberance. 

The 2008 debt crisis taught us that financial markets sometimes screw up in a major way. In my forthcoming book, “Prosperity for All: How to Prevent Financial Crises”, I explain what went wrong and how to fix it. We must  create an institutional mechanism to stabilize swings in the financial markets that always, eventually, end badly.

The second problem is chronic.  Doctors swear to “do no harm”. Economists swear to “block no trade”. That's been a healthy doctrine for most people in the world: if you don’t believe me, just ask the 1.5 billion Chinese workers who saw  growth rates of more than 10 percent a year for decades. And it's been a healthy doctrine for the wealthy and educated elites in the US whose income is derived from owning physical or human capital. But 1.5 billion Chinese workers do not vote in US elections. And American workers have not fared so well. 

The second problem for western democracies is to find a way to redistribute the gains from global trade to working class and middle class voters who have seen their living standards eroded at the expense of Chinese workers and the educated western elites who control the debate. 

Its the wiggles not the trend

Chart 1: Number of unemployed people and number of people in the labor force. Both series normalized to December 2007=100.

A commentator on my blog asks if the reason that unemployment appears to be more important than labor force participation as a cause of recessions is that the scales are different on Chart 2 from my previous blog post. I don't think that's it.

Chart 1 (left) shows the data in a different way. The red series is the civilian labor force. The blue line is the number of unemployed people. Both series were originally measured in thousands of people. The data on the graph have been normalized by constructing index numbers.  In each case I normalized the series in December 2007 (the start of the Great Recession) to 100.

Both series have un upward trend. That's because of population growth: there are more people in the labor force each year and there are more unemployed people each year. My point has nothing to do with the trends. Its about the wiggles. 

Look at it another way. If I gave you the red series and asked you to predict the number of people in the US labor force in 2015, using data through 1990, you'd probably do a pretty good job using lagged values of the labor force, lagged values the US population, and a quadratic time trend. If I also told you there was going to be a major recession lasting from December of 2007 through June of 2009, it wouldn't have helped you much, if at all, in your prediction. Recessions are all about the wiggles. Participation is all about the trend.

It's the Unemployment Rate Stupid

The recent labor market data was abysmal; 38,000 new jobs created against an expectation of 150,000 or more. Unemployment fell slightly to 4.7%, in part because the labor force continued to shrink. A lower unemployment rate is good news and, historically, 5% unemployment has been considered close to full employment. But is that the whole picture?

Chart 1: The Employment to Population Ratio

Some have claimed that we should be looking at the employment to population ratio instead of the unemployment rate. The performance of that statistic is not encouraging as is evident from Chart 1 which shows that the current employment to population ratio is 59.7%, a value that we last saw in May of 2009. That looks bad. But should we be worried? 

In the dim distant past, economists focused on the unemployment rate as a measure of labor market strength. In the 1970s, a group of economists from the real business cycle school persuaded us that unemployment is the wrong measure and that we should look instead, at the hours workers by a representative person. That measure varies for three reasons. Changes in the length of the work week, changes in the  labor force participation rate, and changes in the unemployment rate. The length of the work week for the average person has fallen steadily since WWII and although it goes up and down over the business cycle, the variation in average weekly hours in not the major reason for recessions. What about the labor force participation rates? 

Chart 2: Unemployment and Labor Force Participation

According to a popular narrative, labor force participation is low because of 'discouraged workers'. These are people who would really like to be working but who have dropped out the labor force because they cannot find a job. What is the evidence for the 'discouraged worker'? In my view; not much.

Chart 2 shows male and female labor force participation rates alongside the unemployment rate. Try as I may, I cannot get excited about the changes in the labor force participation rates. They display secular trends that are largely explained by the aging population and by the sociology of an increase in female labor participation in the 1950s and 1960s.

Recessions are not, as some economists have argued, times when workers choose to voluntarily enjoy additional leisure.  They are times when it is hard to find a job because not enough people are choosing to spend their income as opposed to saving it. If the discouraged worker effect were an important explanation of the slow recovery from the Great Recession, we would expect to see significant variation in labor force participation rates at business cycle frequencies. The fact that we don't suggests, to me, that the dismal employment to population ratio is caused by the end of the baby boom. As I argue in my forthcoming book, Prosperity for All: How to Prevent Financial Crises, this is not something that monetary or fiscal policy can, or should, be used to 'cure'. 

_____________

The first version of this post incorrectly posted the employment to population ratio for 25-54 year olds. I have updated it with the civilian employment-population ratio.

Kitten on the Keys

In my misspent youth, I considered a career as a ragtime guitar player.  Two of my idols were David Laibman and Eric Schoenberg who transcribed Scott Joplin for guitar. I sold all my vinyl many years ago (big mistake) and have slowly been replacing it with Amazon digital downloads.

Laibman and Schoenberg have only one album of note, the New Ragtime Guitar. But what an album! One of my favorite pieces is Kitten on the Keys and at one point, I learned how to play the Laibman-Schoenberg transcription of Dill Pickles Rag. 

Why, I hear you ask, would I post this on an economics blog. Because, in googling Laibman and Schoenberg to find out what had happened to them, I learned for the first time, that David Laibman is an economist who retired a few years ago from CUNY in New York. Here is a  link to his page. 

Eric Schoenberg turns out to be his cousin and he sells high end guitars in Tiburon, near San Francisco. Next time I'm in the Bay Area I plan to check out the store. Who knows, maybe I'll get a discount on the $76,000 1938 Martin in exchange for the publicity. 😎