Artificial unintelligence

In recent weeks I have come across several cases in which officialdom appears to be entirely set on sabotaging both their own goals as well as any trust that non-bureaucrats might have in government policy. I will highlight two examples in this article.

First is the UK’s Vaccine Compensation Scheme. The logic of establishing such a scheme is simple but important. Putting aside the merits and disadvantages of specific vaccines, we may agree that in some cases the potential benefits of vaccination greatly outweigh the potential risks. That does not mean those risks are zero, but that, treating individuals anonymously, the total harm that is prevented by vaccinating the population as completely as possible exceeds the total harm caused by the vaccine.

As a relatively uncontroversial example, consider the polio vaccine. Many of us were vaccinated as children or adults. Either we or our parents can remember when the risks associated with exposure to polio, especially for children, were a vivid concern that was ever-present during polio outbreaks. These fears have largely disappeared since the two primary forms of polio vaccination were developed – the inactivated polio (Salk) vaccine and the oral polio (Sabin) vaccine. However, there is some evidence that in a very small proportion of cases both vaccines, but particularly the oral vaccine, can be linked to the onset of paralytic polio or other serious syndromes.

Without question, the social – and individual - benefits of polio vaccination have been enormous. However, those making decisions on whether to vaccinate their children or themselves are not concerned about population statistics but about the risks to their family. The role of a vaccine compensation scheme is, in effect, to say: if the worst happens, though we don’t believe it will, we will ensure that anyone harmed is looked after for as long as necessary and without having to prove negligence or go through lengthy and expensive litigation.

For any economist such schemes are signalling arrangements. They express a public confidence that the benefits of vaccination greatly outweigh the potential harms that is backed by a commitment to fund compensation to those who have the extreme bad luck to suffer harm. For this signal to be trusted, the compensation must be both generous and easy to access, if it is possible to establish a reasonable link between the harm that has occurred and the prior vaccination.

One might have thought that any bureaucracy charged with implementing the Vaccine Compensation Scheme would understand this logic and operate appropriately. Unfortunately, this is where the artificial unintelligence of bureaucracy kicks in. Consider the case, discussed by Norman Fenton in his Substack “Where are the Numbers?”, of the person whose compensation claim for a severe stroke that occurred less than 20 minutes after receiving a booster vaccination for Covid-19 was denied “on the balance of probabilities”.

The sheer stupidity of this decision is hard to grasp. We should be clear. There may be grounds for questioning the biological link between a Covid booster vaccination and a stroke affecting an apparently healthy person. The occurrence of strokes may be (partially) random and very low probability events – i.e. a stroke unrelated to a vaccination 20 minutes earlier – do occur.

However, this is not the issue. Even if the two events were independent, a signalling arrangement only works if the general population believes it. What matters is not some sophisticated professional judgement, but how the signals are understood. That means that compensation must be paid in all cases for which the public – not doctors or statisticians – believe that it is more likely than not that the vaccination caused the harm which occurred.

If claims that seem clear-cut to a reasonable outsider are denied, then the signal given is not the one intended but the exact opposite. Instead of conveying confidence that the benefits of vaccination greatly outweigh potential harms, the bureaucracy is saying, though probably unintentionally, that it fears the cost of compensation will be high because the harm caused by vaccination is large.

Since this should be well-understood, it is hard to be polite about the senior officials responsible for administering the compensation scheme. Their defence may be that they are required to be careful in spending public money. However, that is to forget that the core function of the vaccine compensation scheme is to build up and maintain public confidence in vaccination policies. Being absurdly parsimonious is a classic case of throwing the baby out with the bath water.

My second example is a larger, though less well-known, example of the same behaviour. It concerns what are known as R&D tax credits for UK corporation tax. The core problem is how to encourage businesses, especially start-ups and SMEs, to invest in developing and commercialising new technologies and intellectual property. In the standard model, companies spend money on R&D which is capitalised and written down when and if the products and services are offered for sale.

This means that, for example, a biotech company developing a new diagnostic device must have sufficient capital to fund the company through an extended R&D phase including trials and regulatory compliance. Many run out of money and are obliged to sell their IP to larger companies, often based outside the UK. This is where well-designed tax credits may assist such firms.

The essential idea of all tax credits is that the ultimate recipient can reduce its corporation tax or profits tax bill relative to what it would have paid without the tax credit, thus providing an incentive to do more of whatever the tax credit is linked to. However, the benefit for any company that does not pay any profits tax is small. The tax credit adds to the company’s tax losses which will reduce future payments of profits tax when or if the company earns sufficient profits to offset its accumulated losses.

There are two ways of trying to ensure that companies that do not pay profits tax can benefit from tax credits. The approach favoured in the US is to make tax credits transferable. The federal government has relied heavily on a combination of production and investment tax credits to subsidise electricity generation from solar, wind, geothermal and other renewable energy plants. Most of these plants are operated by special purpose entities (SPVs) which, initially at least, have limited or no taxable profits. However, because their tax credits can be transferred to either associated companies or financial institutions, various financing arrangements can be put in place by which the SPV receives cash to fund its investment and operational expenses in exchange for tax credits that reduce the recipients tax bill. Such deals reflect what is effectively a market in tax credits.

In the UK, HMRC has resisted the idea of making tax credits freely transferable because of the potential loss of revenue. Instead, companies which do not make taxable profits (before taking account of any tax credits) can cash out their tax credits by foregoing tax losses and receiving cash from the government. However, like all tax authorities HMRC is distinctly unenthusiastic about the idea of paying money to companies, so it has hedged the eligibility for receiving money in exchange for tax credits with an increasing number of restrictions. Further, the net cash benefit to the recipient has been greatly reduced in recent years.

The original scheme of R&D tax credits for SMEs was introduced in 2000 but attracted relatively little attention and use until the early 2020s because other sources of funding for small companies were more attractive. Specialist accounting firms started to market the idea of claiming R&D tax credits more actively and the amount of money being reclaimed started to increase substantially. In response, from 2022 onwards the Treasury has announced a sequence of measures that have drastically reduced the value of R&D credits, especially to loss-making SMEs.

The maximum effective benefit that such companies can receive has been reduced by about 50% from 33% of the eligible expenditure before 2023 to 16% from 2024 onwards. The rules on the types of expenditure that qualify for tax credits have tightened and the checks required before tax credits are granted have become much more onerous. In practical terms, what was conceived as a scheme to support research and development across a wide range of activities has been transformed into a scheme to support research, but not really development, in a limited set of science-based sectors.

It is easy to understand the concern of the Treasury about the rapid increase in the utilisation of R&D tax credits. It is integral to the nature of Finance Ministries to hate schemes that threaten to erode the tax base. Politically, it is difficult for any government, especially ones that are desperate to stimulate economic growth, to admit that they are limiting or reducing support for R&D. What happened was a gradual subversion of the scheme by lowering the value of tax credits and imposing increasing onerous administrative requirements on companies that want to claim them, especially via cash payments. Since large companies can make a significant fuss, the restrictions will inevitably fall most heavily on SMEs and, especially, start-up companies.

This brings us back to the ineffable stupidity of bureaucratic organisations – in this case HMRC and the Treasury. The UK is, notoriously, a country that has a strong tradition of scientific research but a rotten record of converting scientific knowledge into viable commercial products, i.e. the development part of R&D. So, what does the Treasury do? Impose increasingly stringent restrictions on the eligibility of development expenditures for R&D tax credits. It ignores a wealth of evidence accumulated over several decades about the failure of the British economy to sustain sophisticated manufacturing activities based on the continuous development of existing and new products.

Why? I would suggest that two factors are particularly important. First, the government is enamoured of biosciences and equally afraid of the threats made by large pharmaceuticals companies to relocate away from the UK. It is easy to pretend that R&D expenditures in biosciences are research, not development. Second, officials have little understanding of small and medium businesses which develop products and services, other than those which boast their “green” credentials. This is part of the longstanding British elite prejudice against the Mittelstand that persists even though these have played a crucial role in contributing to past economic growth.

All tax authorities are suspicious of tax incentives that appear to be too successful, since they are convinced that these are simply mechanisms for tax avoidance. Indeed, such incentives are often presented to potential beneficiaries as tax avoidance arrangements. The difficulty is that by being so suspicious about the motives that prompt businesses to take advantage of R&D tax credits the Treasury and HMRC have responded by sabotaging the original purpose for which the arrangements were introduced. Again, baby and bathwater.

What these two cases illustrate is that policymakers and officials seem to believe that the public is incapable of distinguishing between appearance and reality. That a vaccine compensation scheme and R&D tax credits exist is what matters, not how they perform. Whether any of the expected recipients of such support benefit from the schemes is of little importance. Yet, most people have an acute sense of the difference in public policy between what is claimed and what happens in practice.

Every time there is a report that someone has been refused vaccine compensation on what seem to absurd grounds, or that tax credits are refused or clawed back for not complying with increasingly strict rules, the impression is given that no policy initiative, however meritorious, can be trusted.

The overriding problem for the British state in 2025 is political and government over-reach – the pretence that many issues can be addressed by legislation and official schemes for which there are neither the resources nor the willingness to implement properly. There are sound – even very strong – reasons for both a vaccine compensation scheme and tax incentives for R&D. The stupid mess that officials have made in both cases illustrates a political obsession with appearance combined with a high degree of administrative incompetence.