Tax Analysts®Tax Analysts®

My Subscriptions:

Featured News

January 9, 2012
Start-Ups, Not Small Businesses, Are Key to Job Creation

Full Text Published by Tax Analysts®

By Martin A. Sullivan --

In an October 23, 2011, op-ed in The New York Times, Jared Bernstein neatly laid out some earthshaking facts. The former economic adviser to Vice President Joe Biden summarized recent research that says small business is not the engine of job creation that politicians and the press repeatedly claim it to be. "To the extent that size matters at all for job growth, it's really about new companies that start small," he wrote.

Underlying Bernstein's claims is the work of current and former Census Bureau economists who found "no systematic relationship between net [employment] growth and firm size." (See John Haltiwanger, Ron Jarmin, and Javier Miranda, "Who Creates Jobs? Small vs. Large vs. Young," National Bureau of Economic Research working paper 16300, Aug. 2010.)

The Bush tax cuts are scheduled to expire at the end of 2012. Whether they should be extended for high-income taxpayers will be a hotly debated issue in Congress and on the campaign trail. Central to Republican support for across-the-board extension is the claimed detrimental effect high-bracket tax increases will have on small businesses and job creation. The research cited by Bernstein blows the Republican argument out of the water.

Why, you ask, has this game-changing research not gotten more prime-time attention? The main reason is that no politician of any persuasion wants to do anything but praise small businesses. And the media has no real interest in penetrating beyond the politicians' consensus. It frames all debates in terms of right versus left. If conservatives and liberals agree on a policy conclusion, the media largely accept it as being factually correct.

Republicans seeking tax cuts from rich job creators are just fine with that. But note that even the Obama administration is complacent. It will never attack the conventional wisdom about small businesses and job creation -- even though it would greatly bolster its case for allowing high-bracket tax cuts to expire -- because it does not want to offend the powerful small business community. And that's almost certainly why Bernstein made those arguments only after he left the administration in May 2011.

There is another reason why Bernstein's op-ed has not gotten more attention. You probably noticed the words Biden and New York Times in the first paragraph. Those associations instantly create enough suspicion for any casual reader to discount the results as merely one side of a partisan debate. Sure, Bernstein, who is now employed by the left-leaning Center for American Progress, is not losing any friends by advertising the research of Haltiwanger and his colleagues. But there is nothing to suggest political bias is driving the results. For one thing, Haltiwanger has been described by Washington Post columnist Steven Pearlstein as "a straight shooter, without a trace of ideological bias." (See "A Healthy Dynamic in Job Creation: Destruction," The Washington Post, May 31, 2011.)

Then there are the conservative economists who agree that small businesses are not special when it comes to job creation. Veronique de Rugy, formerly at the American Enterprise Institute (AEI) and now at the Mercatus Center at George Mason University, agrees with Haltiwanger. "The conventional claim about small businesses' job creating rests mainly on statistical fallacies and misunderstanding of the data," she says ("Are Small Businesses the Engine of Growth?" AEI working paper 123, Dec. 8, 2005). De Rugy cites conservative icon Milton Friedman (Friedman, "Do Old Fallacies Ever Die?" J. of Econ. Lit., Dec. 1992). She concludes that "the claim that small businesses are the fountainhead of job creation does not hold water." Similarly, Alan Viard and Amy Roden of AEI write that "the most common argument for preferential treatment of small business -- its uniquely powerful role in job creation -- does not stand up under scrutiny. . . . Careful statistical studies do not assign any special role to small firms." (See "Big Business: The Other Engine of Economic Growth," June 23, 2009, available at

The work of Haltiwanger and his colleagues deserves more attention because it is standing up to the scrutiny of other professional economists. In other words, it appears that it will be one of those economic claims subject to endless debate. For example, University of Chicago economists Erik Hurst and Benjamin Pugsley build on Haltiwanger's research ("What Do Small Businesses Do?" NBER working paper 17041, Oct. 2011) and join the emerging consensus view:

    Often subsidies targeted at increasing innovative risk taking and overcoming financing constraints are focused on small businesses. Our analysis cautions that this treatment may be misguided. We believe that these targets are better reached through lowering the costs of expansion, so they are taken up by the much smaller share of small businesses aspiring to grow and innovate.

The Kansas City, Mo.-based Kauffman Foundation has devoted a considerable portion of its $2 billion in assets to studying and promoting entrepreneurship. Based on research similar to Haltiwanger's, the foundation's numerous publications repeat the finding that new firms -- not small firms -- are the source of almost 3 million net new jobs that on average are created in the United States each year. "Virtually all of the growth in U.S. jobs has been driven by the formation of firms less than five years old," the foundation reports on its website. "These new firms have been disproportionately responsible for commercializing the cutting-edge innovations that characterize modern life." (Based on its findings, the foundation developed and proposed the Startup Act. Formal legislation was introduced in the Senate by Sens. Jerry Moran, R-Kan., and Mark R. Warner, D-Va., on December 8. It includes a permanent exemption for start-up capital gains and tax credits for investment qualified small businesses.)

If none of that convinces you that the new research destroys widely held notions about small businesses and job creation, perhaps the absence of substantive criticism from small business advocates will. On their websites, the Small Business Administration (SBA) and the National Federation of Independent Business (NFIB) stick to the conventional wisdom and simply ignore the new findings that only a subset of small firms -- the start-ups -- drive job creation. They continue to lump all small firms together. (See NFIB, "Speech Material," available at "Small business has created about two of every three net new jobs in the United States since at least the early 1970s." And see SBA, "Frequently Asked Questions," available at "Small firms accounted for 65 percent . . . of the 15 million net new jobs created between 1993 and 2009.")

In support of its claim, the SBA cites a study it funded that emphasized the importance of mature firms in job creation (Zoltan J. Acs, William Parsons, and Spencer Tracy, "High-Impact Firms: Gazelles Revisited," June 2008, available at But that study did not take into account the large role played by start-up firms. (See the critique by Scott Shane, "Entrepreneurial Job Creation Statistics Are an Economic Rorschach Test," Mar. 15, 2010, available at Obviously, the SBA and the NFIB will have a hard time accepting research showing that most mature small businesses are not job creators but job losers.

Statistical Illusions

Usually Tax Analysts likes to spare readers the details of economic research and get straight to the bottom line. But sometimes -- as during the great debate on capital gains in the 1990s -- economic methods move to center stage. In these cases, the passage or defeat of major tax legislation can hinge on obscure assumptions made by obscure economists. Because the role of small businesses will be central to debates on both job creation and the extension of the Bush tax cuts, some readers will want more detail on how, in light of all the recent research, politicians and small business advocates can continue to claim there is a special role for small business in job creation. The following is a quick overview of the key statistical issues.

(1) Volatile net job creation. Suppose the United States exported $1 billion of goods to a small country from which it imported nothing. And suppose further it had a zero trade deficit with a large country from which it imported and exported $100 billion of goods. If we look only at net figures (plus $1 billion of trade with the small country and zero for the larger country), we might be fooled into thinking that trade with the large country was unimportant. After all, trade with the small country represents 100 percent of net exports. And therefore shouldn't we focus our policies on increasing trade with the small country?

The same statistical trap beguiles analysis of job creation data. Large amounts of job creation and destruction underlie relatively small net changes. For example, 2.2 million net new jobs were created in the United States in 2005. That is the result of 17.9 million new jobs created against 15.7 million lost ( Especially in years when economywide job creation is small, it is easy to say that any sector with net job creation was responsible for a large share of overall job creation. Mostly because of start-ups, small businesses are routinely net job creators, and that amount of net job creation can be large compared with the total. That is the source of the claims of the SBA, the NFIB, and most politicians. But there are massive swings in gross job creation and gross job destruction by both large and small business. Small changes to any of those large movements could play a critical role in increasing U.S. net employment. The arithmetic may be correct, but net employment growth figures leave a false impression. De Rugy nicely describes it as "netting out reality." Analysis based on net employment data effectively treats large portions of the job market as nonexistent. Therefore, it can be dangerously misleading if it is used as the basis for policy.

(2) Regression-to-the-mean fallacy. This problem, common in statistical analyses looking at data over time, results in an inverse relationship between size and growth (in this case, a relationship in which small firms grow faster than large firms) even when they are independent of each other. Friedman has referred to it as "the most common fallacy in the statistical analysis of economic data." It arises because there are always transitory errors in measurement of data and temporary noise that has nothing to do with the issue under study. Therefore, firms that are characterized as small on average are really larger than they appear in the data, and firms characterized as large are really smaller than they appear. When these temporary factors fade, firms regress to average size. And so, over time, small firms on average appear to grow faster than large firms. When economists correct for this bias, they find the amount of small firm job creation is not as large as the unadjusted data suggest. And in some industries, small firms have no extra job growth at all (Haltiwanger et al. cited above; and David Neumark, Brandon Wall, and Junfu Zhang, "Do Small Businesses Create More Jobs? New Evidence for the United States From the National Establishment Time Series," Rev. of Econ. and Stat., Feb. 2011).

(3) Left-out variable bias. As an example, suppose that success in a particular business is entirely explained by IQ. And suppose, as is entirely likely, that a researcher without data on IQ finds a strong relationship between years of education and business success. The omission of a key variable leads to the incorrect conclusion that more education can increase success. If business recruiters naively used those results, they could end up hiring unproductive, overeducated workers, and the government could end up subsidizing education that does not contribute to productivity growth.

In the past, economists did not have a lot of data on firm age. Start-ups, young firms, and mature firms were all lumped together. So when economists looked at the data, they could see that small firms had higher rates of job creation than large firms, but they could not see beyond that. But over the last decade, the Census Bureau has developed a new data set it calls Business Dynamics Statistics. These data allow economists for the first time to simultaneously examine firm size, firm age, and employment growth on an economywide basis ( It is analysis using this new data set that has put the nail in the coffin of the conventional wisdom. Youth, not smallness, is the key to job creation. Start-ups, not mature firms, are the engines of job growth. Policies that provide incentives to all small businesses are poorly targeted because only start-up firms are net job creators. Mature small firms are net job losers.

Not Harmless

Like education, homeownership, motherhood, and apple pie, small business enjoys tremendous public and political support. If government is kind to these likeable economic entities, is there really any harm? Well, yes. Economics is coldhearted. Subsidies should not be the prizes of popularity contests. Without sound justification (such as correcting a market externality), government interference with resource allocation will hurt rather than help growth.

In a 1995 paper, Douglas Holtz-Eakin, adviser to Sen. John McCain during his 2008 presidential campaign, wrote that "subsidies to small firms constitute a 'tax' on growth. . . . [As] the firms' success and capital needs grow, the tax code raises -- other things being equal -- the cost of capital." (See "Should Small Businesses Be Tax-Favored?" Nat'l Tax J., Sept. 1995.) Hurst and Pugsley also point out that if benefits are eliminated for firms with growth above a specific size, our "policies, as currently structured, may actually inhibit firm growth." De Rugy calls this a "perverse incentive" that will "lead to the misallocation of resources away from their most productive uses and will interfere with the natural growth and evolution of firms."

Viard and Roden also warn against directing tax benefits to small businesses: "Even if small firms play no unique role in job creation, one might think tax policies favoring small firms over big ones are harmless. Unfortunately, that is not true. Preferences for some firms over others interfere with the market's allocation of resources and disrupt the efficient workings of the economy."

In their research, Hurst and Pugsley find that most small firms do not grow or innovate. So they warn that tax and other subsidies for small businesses can hurt growth: "To the extent that these subsidies alter the behavior of firms who start for reasons unrelated to growth and innovation, the policies can be distortionary by allocating more resources to the small business sector than would otherwise be optimal."

All this tells us that perhaps the best economic case for keeping the top individual rate at 35 percent is the benefit lower rates would provide to all sectors of the economy. But that bland economic reasoning does not connect with voters the way a story about struggling small business does. Republican politicians will never make that mistake. That would be surrendering their most potent weapon in the battle to extend the Bush tax cuts. Better to upset a few economists than lose the support of the public, which has been taught to equate tax increases for the wealthy with tax increases for small business job creators.

About Tax Analysts

Tax Analysts is an influential provider of tax news and analysis for the global community. Over 150,000 tax professionals in law and accounting firms, corporations, and government agencies rely on Tax Analysts' federal, state, and international content daily. Key products include Tax Notes, Tax Notes Today, State Tax Notes, State Tax Today, Tax Notes International, and Worldwide Tax Daily. Founded in 1970 as a nonprofit organization, Tax Analysts has the industry's largest tax-dedicated correspondent staff, with more than 250 domestic and international correspondents. For more information, visit our home page.

For reprint permission or other information, contact