Robert Chote: Boffins becalmed: we're good at inventing incentives, but not at producing inventors

Gordon Brown promised last month to "create the best incentives and make Britain the best place for R&D". Promoting innovation may well be an important theme in his pre-Budget report on Thursday - it usually is. But while it is always tempting to propose new policy initiatives, it is also important to take a long, hard look at those that are already in place.

For many years, the Department of Trade and Industry presided over a bewildering range of business support schemes, some of which were designed to encourage innovation. It has pruned them back significantly but political pressure to go further is growing, with the Tories now promising to shrink the DTI and the Liberal Democrats to abolish it.

In assessing innovation policy, three questions need to be answered. Do we have an innovation problem in the UK? If so, why? And do we have the right policy instruments to tackle its causes?

Whether innovation performance is adequate or not is hard to judge from first principles, as there is no obvious way to determine the ideal level of research and development. We should be concerned with both the quality and quantity of R&D, but the policy debate typically begins with international comparisons of how much is spent.

So how does the UK shape up? We spent 1.9 per cent of national income on R&D in 2001, compared with 2 per cent across the EU and 2.8 per cent in the US. Two years ago, the European Council set a target of 3 per cent for the EU by 2010, while the Government here has the less ambitious - but perhaps still over-optimistic - goal of reaching 2.5 per cent by 2014.

Differences in R&D intensity in part reflect differences in industrial composition. Take France and Germany, which both spend more on R&D than the UK. A new study by Laura Abramovsky, Rupert Harrison and Helen Simpson at the Institute for Fiscal Studies attributes two-fifths of Germany's lead in R&D over the UK in 1999 to its larger motor industry (typically R&D intensive) and smaller service sector (typically less R&D intensive).* Differences in industrial composition explain less of the gap with France.

Looking within sectors, the study found that the UK had higher R&D spending than France and Germany in services and pharmaceuticals. But this was more than outweighed by lower R&D spending in most of the rest of manufacturing, especially communication equipment. (It is hard to make a similar comparison with the US as some of the R&D attributed to manufacturing in Europe is attributed to services in America.)

Further evidence that the UK lags leading European competitors can be found in the Community Innovation Survey, which asked manufacturers whether they had introduced any innovative new products or processes between 1998 and 2000. The proportion of such firms in the UK was lower than in France and Germany, and about level with Spain.

Interestingly, the proportion of innovative companies in Britain engaged in collaborative research or transferring technology from universities, research institutes and the like was broadly in line with the numbers in France and Germany. So the problem is not that innovative firms fail to make these links, but that firms in the UK are less likely to be innovative in the first place.

Firms might be expected to carry out too little innovation for three main reasons - which might, therefore, justify government action:

First, R&D may have beneficial spillovers for other companies, and for society, that the firm undertaking the investment cannot capture. Tax breaks and grants can be used to encourage them to undertake more R&D than they would otherwise.

Second, innovation that requires several firms to collaborate may not take place if some find it hard to make a commitment. Co-ordination problems are particularly likely if the firms need to sink a lot of money into an investment specific to the project. So it may be sensible for government to subsidise these costs or underwrite some of the risk.

Third, firms may be unaware of potential research partners or a particular technology. Companies should be willing to pay for such information, but they will not if they do not know it is there. Government may be able to help here.

It is hard to know whether the lack of innovative companies in the UK demonstrates that these problems are more pervasive here, or that government policy is less effective in getting firms to make the leap. Maybe both. Some evidence suggests that British firms find it more difficult than German ones to get sufficient finance to allow them to start engaging in R&D.

Several schemes are designed to encourage research and development in the UK. The R&D tax credit is the biggest, costing around £430m last year. Other schemes, funded by the DTI, include Grants for R&D, the Collaborative Research and Development scheme, Knowledge Transfer Partnerships, the Knowledge Transfer Network and the Higher Education Innovation Fund. These and other such initiatives cost around £300m. The Government devotes much larger sums to direct support for science and regional policy, which can provide help too.

One obvious observation is that many more schemes exist than rationales for intervention, which suggests we should be wary of unnecessary overlap. For example, the justification for the Grants for R&D scheme is less powerful now that we have the R&D tax credits.

As you would expect, fewer service sector firms take advantage of public support for innovation than manufacturers. But given the recent strong growth of R&D in services and the importance of that sector to the UK economy, it is worth asking if these schemes do enough to meet its needs.

The Government's recent rationalisation of its pro-innovation schemes is welcome, but it is important to keep them under review and expose them to rigorous evaluation. If the system could be made simpler and more effective in encouraging first-time innovators, so much the better.

Robert Chote is director of the Institute for Fiscal Studies.

There's not enough give in the golden rule

The political significance of the pre-Budget report will centre on the health of the public finances, the fate of Gordon Brown's "golden rule" and whether this means taxes will have to rise after the election.

The story so far: seven months into the financial year, the current budget deficit - the gap between tax revenues and non-investment spending - has not shrunk as the Chancellor had hoped in March. If this continues, the deficit could be £12bn higher than expected.

The golden rule requires the current budget to be in surplus or balance over the ups and downs of the economic cycle, so the Government only borrows to invest. The Treasury says we are now in the penultimate year of a seven-year cycle beginning in 1999-2000.

At Budget time, the Treasury predicted that it would come in on the right side of the golden rule by about £8bn. The projected overshoot in borrowing this year would more than wipe this out, although it would be a closer call if Mr Brown manages not to spend the £3bn contingency reserve in the spending plans for this year and next.

But the Chancellor is adamant the rule will not be broken. His officials expect oil revenues to boost tax receipts and spending growth to slow. He could well be right - the public finances are notoriously hard to predict.

Just to be on the safe side, he could also choose to ignore the deficit pencilled in for 2005-06 by arguing that spare capacity in the economy has been used up and that we are therefore already in the last year of the cycle. This seems in line with the Bank of England's view. It would make the rule easier to meet in the short term, but would strengthen the case for announcing fresh tax increases after the election.

The hoo-hah over whether the rule will be met in the short term is certainly entertaining for politicians and the press, but the Chancellor's insistence that hell will freeze over before he breaks it has raised the political stakes far in excess of its real economic importance.

The golden rule - unlike the inflation target - is asymmetric. So the test is pass/fail, rather than get the ball as close as you can to the goal. Pursuing an asymmetric target when you inevitably face large forecasting errors requires a mature debate about the acceptable risks of missing it. Alas, the political environment makes that highly unlikely.