Should We Support Small Businesses?

Are they really the job creators they’re cracked up to be?

The US government has a host of programs to support small businesses. Small businesses are eligible for subsidized loans from the Small Business Administration, which spends billions a year for this purpose. A portion of federal contracts are set aside for small businesses, with the Biden administration being extremely proud to have given 178 billion in contracts to small businesses in 2023. You have also seen countless instances of romanticizing small businesses – we are called upon to “support your local small business”, to shop small, and that small businesses create most jobs in America. This messaging is so ubiquitous that I have little interest in looking up examples, and leave it to the reader to supply them.

All of this is unbearably wrong. Small businesses are consistently worse in every dimension one could care about. Though comparing firms is difficult due to them selecting into different products and employing different inputs, they are less productive, pay their workers less, and are worse managed.¹ The large firm wage premium has been falling over time, partly due to manufacturing becoming a smaller part of the economy, partly due to very successful and high-paying companies (like startups, and partly due to very large firms no longer having the premium they once did; nevertheless, wages are higher at larger firms. Size of firm strongly predicts management quality. (See especially the work of Bloom, Sadun, and Van Reenen, well summarized in “Why Do Management Practices Differ across Firms and Countries?”). Turning to overseas evidence, the persistence of small firms is often due to explicit policies favoring them. Removing those handouts, as in Martin, Nataraj, and Harrison (2017), causes employment to increase. Relatedly, improvements in the quality of the judicial system (which allows for firms to be run through contracts, and not entirely by family members) show a marked tendency for the size of firms to increase.

The primary competitive advantage of small businesses is that they can evade taxes and regulations. Indeed, most small firms are just vehicles for deducting personal consumption as “investment”. David Leite (2024) uses evidence from Portugal to show that firm owners shift about a third of their consumption over to the firm. Most of the earnings of business owners are actually disguised labor earnings, as profitability tanks after the owner dies or retires. We exempt many small firms from regulation – which is, admittedly, fully justifiable, as regulation is often a fixed cost. Nevertheless, this does tip the scales toward making firms small, and prevents the successful ones from growing. Trebbi and Zhang (2022) show an inverted U-curve for the cost of regulatory compliance, with the smallest firms paying very little. And of course, one pays nothing at all if the point of being small is not just to questionably declare expenses, but dodge taxes entirely. In the context of Italy, Pelligrino and Zingales (2019) argue that the great mass of small, unproductive firms were unable to afford investing into IT, leaving Italy poorer than otherwise.

Small businesses are not especially important creators of jobs. New businesses are. These start small, but the successful ones grow. Haltiwanger, Jarmin, and Miranda (2013) is the key reference on this, as before the Longitudinal Business Dataset we lacked the tools to answer this. Once you control for firm age, there is no relationship between job creation and firm size.

A small detour. Politicians like claiming things like “80% of jobs are created by small businesses”. However, the method they rely upon to claim this would be equally able to say that small businesses are responsible for 100% of new jobs, or 150%. It literally is just taking the total number of new jobs in small businesses, and dividing by the net change in jobs in the economy as a whole. You completely ignore job losses. (I learned about this from De Rugy (2005) “Are Small Businesses the Engine of Growth?”.)

Instead, most small businesses have no interest in introducing new products or techniques. There is a divide between “subsistence” and “transformative” entrepreneurs. Subsistence entrepreneurs are not trying to do anything really new, but value something about the form of owning one’s own business, such as the control over their own hours. This is the great majority of business owners. These are everything from plumbers to accountants to mechanics to restaurants to gas stations to insurance agents. Hurst and Pugsley (2011) show that these firms, after being founded, do not grow at all. Only 34 percent of firms report that they were starting a business to generate income, and only 40 percent started because they had a good business idea or wanted to create a new product.

Transformative entrepreneurs are start-up founders and the like, and differ insofar as they employ people engaged in research and development. If the government blindly shovels money at small businesses, then all it is doing is taking from those who create ideas to small scale rentiers. To the extent that the government should subsidize businesses, they should subsidize those whose technological innovations are more likely to spill over onto other firms. In other words, it should subsidize research, not something which weakly correlates with research!

So, if we believe that it is incorrect for the government to transfer money to accountants and plumbers, there is no reason to support small businesses just because they are small.

1. The biggest problem is employing more or less skilled workers. One can easily imagine cases where the small firms consist of skilled craftsman, and the large firms employ low-skilled workers. It is possible for the wage paid in the large businesses to be lower than the wages in small businesses, but higher for those particular workers if everything were a small business.