The objections raised below suggest that many of the new jobs that would trigger the tax break would be "created anyway." This is certainly true, however even if just one of every six jobs created in 2010 is directly spurred by the credit (as suggested by the analysis above), the increased tax revenues and lower safety net spending that are generated by the new jobs would make its net fiscal cost less than $5,000 per job, according to Bartik and Bishop. This is a bargain when compared to many other options under consideration.

On a political note, it's a sad but true political fact that economic recovery proposals passed by Congress include billions of dollars in direct benefits for business (witness the net operating loss carryback provision in the just-passed UI extension). These rarely do much at all for actual job creation. This time, why not tie some of these business benefits directly to hiring?

The closest analogy to the kind of job creation tax incentive now under discussion is the New Jobs Tax Credit (NJTC), which was enacted in 1977 during the recovery from the 1974-75 recession. The goal of the credit was to encourage job growth by providing a generous subsidy for new hires that increased a company’s payroll above 102% of the payroll level in the previous year. The credit gave subsidies to employers for net employment growth in 1977 and 1978. For companies claiming the credit, the NJTC reduced the average compensation cost of a subsidized worker by about 20% to 25%. In other words, the credit reduced by up to one-quarter the net cost to employers of putting subsidized workers on their payroll. Len Burman in his contribution to this discussion has pointed to recent assessments of the NJTC experience in the TaxVox blog (here and here).

A few academic researchers who’ve analyzed the NJTC are optimistic about its effects on employment growth in 1977 and 1978. I am less persuaded the credit had much effect on employment in that recovery. My interpretation is that the credit added $5.7 billion to the 1977-78 budget defict ($18 billion in current prices) while inducing very slight gains in average employment.

At the beginning of the 1980s I worked in the office of the Secretary of Labor. Congress required the Secretaries of Labor and Treasury to submit a joint report on the impact of the NJTC. An economist in the Treasury Department and I were deputized to begin preparing the report. Although I did not remain in the government long enough to finish the report, I remember our initial findings. The final report to Congress, submitted in 1986, can be found here. (I played no role in writing that report.) For proponents of job creation tax subsidies the findings of the Labor-Treasury study are not encouraging.

Of the business firms which filed tax returns, about 10% claimed the NJTC in 1977 and 19% claimed it 1978. Companies’ take-up of the credit was low for two reasons. First, many businesses, including ones that expanded, did not qualify. This is obviously the case for all companies where 1977 and 1978 employment was below 102% of the level in the preceding year. Even among firms meeting the threshold, however, some were not profitable and hence could not claim the credit. A second reason for low take-up was employer ignorance.  Many eligible businesses were unaware they were entitled to claim the NJTC.   The Labor-Treasury report estimates that about 70% of firms eligible for the credit failed to claim it on their tax returns. Even among firms with more than $10,000,000 in annual sales, 42% of eligible companies did not claim the hiring subsidy. This is a shocking number, because nearly all large firms employ tax specialists to prepare their tax returns. If large firms with employment growth entitling them to a tax credit did not claim the NJTC on their returns, it is hard to believe the credit influenced their decision to expand employment.

Among the firms actually claiming the credit on their returns, a large percentage received no tax subsidy for a marginal new hire. Congress imposed limits on the credit to reduce common abuses and to target most of the subsidy on small and mid-sized firms. One result of these limitations was that many expanding firms quickly reached the maximum subsidy they were permitted to claim; they did not receive any additional subsidy for their marginal new hires. The overwhelming majority of all new employment growth that occurred in the companies claiming the NJTC occurred without any marginal subsidy for the last worker hired. Companies received a tax credit that reduced federal revenues, but the subsidy did not cut the companies’ net hiring costs for the last workers they hired. This strongly suggests companies’ end-of-year employment levels were little affected by the credit.

The problems with tax subsidies for marginal employment increases are well known to analysts. Hiring decisions are often made by company employees who are ignorant of the credit. Limits on the credit to reduce employer abuses will also eliminate subsidies for many employers willing to expand employment on the margin. If the credit is designed to give a subsidy to all new hires, an overwhelming fraction of credit payments will go to subsidizing employment that would have occurred in the absence of the credit. Some of the problems can be minimized in a carefully crafted hiring subsidy program. Most of the problems, however, are unavoidable. How effectively can the government disseminate knowledge about a tax subsidy that may only last one or two years? How do we discourage a company from laying off current workers in order to become entitled to a subsidy for the new hires who replace them? How do we structure the tax credit so employers do not receive subsidies for splitting a single well-paid, full-time job into two poorly paid, part-time jobs? And how do we accomplish all of these goals while keeping the cost of the credit manageable?

There are partial answers to some of these questions. But our experience with the NJTC in the 1970s shows that crafting a well-designed subsidy is not easy. It may be impossible.

For a more recent example, U. Florida professor, Sarah Hamersma, found that the welfare to work tax credit and the work opportunity tax credit have mostly subsidized hiring that would have happened anyway. Indeed, many employers often don’t even know about the credit until after their hiring decisions are made. This suggests that a major education campaign aimed at informing employers and workers about the new program is a minimum requirement for success.

In theory, an incremental subsidy could boost hiring by making new labor cost less than the value of what it produces. But there are a lot of practical problems. For one thing, business credits are typically not refundable. That means that they are of little or no value to companies suffering losses, but these are of course the companies most likely to have slashed jobs. A refundable credit creates issues of its own. For one thing, it costs a lot more. And, would we make the credit available to nonprofits? I'm very troubled by the notion of entities that have opted out of the tax system filing returns only for the purpose of collecting credits. (This may seem like a theoretical concern, but I remember testifying in 1999 opposite the president of Goodwill, who argued passionately that the work opportunity tax credit should be made available to that nonprofit.)

And defining the qualifying incremental job is complex and can create bad incentives, as Greg Mankiw has explained.

There are problems with timing. A company suffering losses, with restricted access to capital markets, is unlikely to spend a credit that won't be received until possibly a year after the hiring is done. To make use of new labor, a company has to be able to sell more, but that might require price cuts during a recession. In that context, it is not just marginal cost that matters. And, the fact is, companies are going to be very reluctant to add to payrolls until they see clear evidence that the economy is recovering. A well-timed tax credit might speed up that process a bit, but it's a clumsy and inefficient mechanism to do that.

A better option would be a temporary cut in payroll taxes, as suggested by Mankiw and many other Republicans. It works for companies suffering losses and nonprofit entities (and most state and local governments, which are also hurting) and would be relatively easy to implement. It would boost companies’ cash flow and raise workers' after-tax income. It should boost spending and would provide a small inducement to additional hiring—and reduce the pressure on struggling companies to eliminate more jobs.

And, it could be a rare bipartisan effort to address what is still a very scary economic situation.