John Haltiwanger is a Distinguished University Professor in the Department of Economics at the University of Maryland. Haltiwanger has devoted much of his career to the study of job creation and firm dynamics, what drives them and how they contribute to growth in jobs, productivity and in the economy generally. The following interview was prompted by a recent draft chapter titled “Job Creation and Firm Dynamics in the U.S.” prepared for an upcoming National Bureau of Economic Research (NBER) publication. In the chapter, John raises his concern about the decline in job creation in startups and small firms and the consequent loss of dynamism in the U.S. economy.

GF: For much of your career you have looked at the dynamics of the economy and the job market to understand what creates job growth. Can you say a little about that work?

JH: The public tends to focus on net job growth. For example, in a good month (which we haven’t had in a while) we would normally add about 200,000 jobs. That means that in a good year we would add 2.5 million jobs. That is a lot of jobs — enough to drive the unemployment rate down on a pretty regular basis.

My work focuses on the fact that we need a lot of dynamism to create those 200,000 jobs. In particular, we need a lot of job creation to offset an almost equal and powerful process of job destruction.

Looking at our “good year” example, we would normally create 18.5 million new jobs. These jobs come from startups, expansion of existing firms or the addition of new jobs in existing firms, many of which may require new skills.

Equally important is the fact that even in that “good year,” about 16.5 million jobs will be destroyed by business failures and existing businesses contracting or changing the mix of their jobs. Thus, behind that 2.5 million net increase in jobs in an average good year is the fact that as many as 34 million jobs might either be created or destroyed.

Given total U.S. employment of about 130 million, we have a “churn” that is roughly equal to about 26 percent of total employment. Over our history, this churning — although it hurts some people and helps others — has been the source of the economic dynamism that has been part of the unique character of the United States.

GF: So what happened during the financial crisis and recession? Did all of the churning stop?

JH: To the contrary, the churning continued, but job destruction exceeded job creation by a wide margin. For example, between March 2008 and March 2009, we lost four million jobs net. This net job loss was the result of having destroyed 19 million jobs (up from a more usual number of 16.5 million, as I mentioned before) and only creating 15 million new jobs. It is worth noting that, even in times as dire as this recession was, we still created 15 million new jobs over this time period.

GF: Where do all these jobs come from — and go?

JH: New firms account for an important share of the new jobs — about 18 percent. The problem with startups is that a lot of new firms fail, so they also account for a large share of the job destruction — about 17 percent. Nevertheless, these young firms are an important part of the “creative destruction” that is a hallmark of a dynamic, innovative economy.

Existing firms create the rest of the jobs, either through new job openings in existing establishments (e.g., an establishment is one of the firm’s locations) or through expanding existing establishments. Unfortunately, these firms lose about as many jobs as they create. Established firms contract or go out of business.

Thus, most or all of the net new job creation comes from young firms and startups. The number of jobs in established firms remains roughly constant over time.

GF: Have these trends changed?

JH: Yes, and I might say, in a rather disturbing way. Chart 1 shows the percent of total employment that is accounted for by job creation and job destruction since 1990. [1] It reveals some important trends that do not bode well for the current job recovery.

Job creation has been headed down in a pretty persistent way since 2000. Job creation rose slightly to a quarterly rate of about 9 percent of total employment in the 1990 recession. Still, job destruction rose even more during that recession. During the 1990s, job creation settled at about an 8 percent rate. Job destruction fell sharply during the 1990s, as did the unemployment rate.

Job creation dipped during the 2001 recession (more normal behavior than in 1991) but the recovery was startlingly weak. Since 2000, job creation has never reached the rates of the mid-1990s. Indeed, in the current recovery, job creation remains at the lowest rate as a percent of total employment of the post-World War II period.

Chart 1: The decline in job creation is a long-term trend

GF: The high rate of job destruction during the recent recession — and long-term trend toward weak job creation — looks pretty alarming.

JH: Frankly, these trends are alarming. Earlier we discussed the annual measures, but Chart 1 shows that, even after the recent recession, job creation remains quite anemic. It is apparent that this reflects both the continued weakness of the U.S. economy in recovering from the recession and the evident downward trend in job creation.

GF: Why is that?

JH: One reason is that the rate of job creation among small startup firms, which is the major source of all net job creation, has slowed down.

Chart 2 shows the gross job creation measured in millions of jobs by size of employer. In the 1990s, firms with 19 or fewer employees created between 10 and 11 million jobs a year. Today, this number is less than nine million — a decline of about 25 percent. Related evidence shows this decline is associated with the decline in startup firms.

More important, there appears to be no recovery. The job creation behavior of these small and often young firms over the past two years mirrors the early part of the last decade. If anything, it is the larger firms that are beginning to create jobs.

In sum, if you look at Chart 3, you can see that the pace of job destruction — the rate at which firms are contracting and laying off workers — has declined dramatically during this recovery; but as Chart 2 shows, the pace of job creation is has not picked up.

Chart 2: Job creation is down not only in startups but also in established firms

Chart 3: The pace of job destruction has dropped sharply during the recovery

GF: What do you think is behind this decline in the trend in job creation?

JH: It is really hard to know. One reason may be uncertainty. If you don’t know where the economy is going, not just in the sense of economic growth, but also how economic opportunities will be distributed among sectors, technologies and even geographies, it is hard to make forward-looking decisions. Certainly the uncertainty about future taxes, regulation and healthcare costs all plays into this mix.

It is interesting that in the early 1980s, after the 1981-82 recession, job creation really took off. Was the business environment of the 1980s more certain? Probably. Still, uncertainty is not the whole reason for the failure of job creation to pick up today.

Demographics may play a role. An aging population is not as ambitious and risk-taking as one that is younger. Startups are typically initiated by people in their late 30s and early 40s. The millennial generation is coming along, but they haven’t arrived at the age yet when people tend to be willing to take the entrepreneurial plunge.

GF: In “Job Creation and Firm Dynamics in the U.S.”, you seem to be not very optimistic about the effectiveness of policy addressing these problems.

JH: A good place for policy to start is addressing anything that affects the business climate, especially reducing uncertainty. Whenever the government attempts to target specific business types (e.g., policies targeted to specific industries or to small businesses), it can backfire. Tax-credit policies and special loan programs usually induce the weaker businesses to stay in business longer.

Over the years, the U.S. economy has gotten a lot out of experimenting with new ways of doing business and producing new products with new processes. Startups play a critical role in this experimentation. An important part of this experimentation is that some experiments fail. They fail for a variety of reasons, and when they do fail, the entrepreneurs and the workers from the business failures move onto other activities. Policy makers usually slow this process down.

GF: This leads me to my last question, and that is how this work links to the unemployment situation we face in this country today. I have had the feeling that the nature of the recession process has changed over the years. Where it was more of a demand management process in the immediate post-World War II period, it is dominated by restructuring of major sectors to a greater degree today.

JH: I hate to end on a negative note, but the prospects for the newly unemployed today are not good. Chart 4 compares the unemployment escape rate to the unemployment inflow rate. Both of these series are derived from the Current Population Survey. These are not numbers the public is as familiar with compared to the unemployment rate.

The unemployment escape rate measures the probability that someone who is unemployed in one month will find a job in the next month — in other words, how quickly someone who loses his or her job will find another. The unemployment inflow rate measures the number of people who become unemployed each month as the percentage of total employment. The red line shows the unemployment inflow rate; the blue line is a three-month moving average of the unemployment escape rate.

Chart 4: People who become unemployed today are much more likely to remain unemployed than in the past

At the end of the 1960s, the share of the employment base that entered unemployment surprisingly was not very different than it is today — about 2 percent, or about one million people. However, the probability that a person would find a job the next month after being unemployed was roughly 75 percent. In other words, people moved fairly easily from one job to another. Even during the aftermath of the 1970s recession and as the numbers unemployed rose, that probability remained at 50 percent.

We can see from this chart that a big adjustment occurred during the 1970s. Once becoming unemployed, the probability of finding a new job quickly dropped not only in recessions but also in recoveries. The time it took for a newly unemployed person to find a new job increased.

That trend took a dramatic turn for the worse in the recent recession. The probability of a newly unemployed person of finding a job in the next month has dropped to 20 percent — one half in some cases of the lows of past recoveries. The effect of these trends can be seen in the sharp rise in the long-term unemployed.

The pace of the movement of Americans out of unemployment is today roughly comparable to the rates we find in Western Europe, where the movement out of unemployment is very slow. Previously, during the 1980s and 1990s, we had distinguished ourselves from Europe with our dynamism, innovation and productivity. We established our uniqueness in part because of the rapid pace of job creation in small and startup firms and the churning of economic opportunities in the U.S. economy.

So, Gail, to go back to the question you pose in the theme of this interview: Is the United States losing its dynamism? I fear that the answer is yes.

The pace of job creation is slowing and, while businesses are no longer contracting as they were during the recession, they are also not responding to new economic opportunities and creating new job opportunities.

To your point, this is not a demand management policy challenge we face in the U.S. economy today. Rather, it is a challenge that goes to the foundations of the slowing pace of dynamic change in U.S. firms themselves, to the climate for entrepreneurship and the willingness of Americans to take entrepreneurial risks, and to the structure of our labor markets generally.

GF: We have a lot of work to do. These are daunting problems. Thank you.

¹ All charts are taken from “The U.S. Labor Market: Still Anemic Two Years ‘After’ the End of the Recession,” a presentation prepared by John C. Haltiwanger, September 2011.