Why don’t policymakers ever learn?

The current mess concerning the future of the Scunthorpe steelworks in North Lincolnshire prompts the exasperated question in the title to this article. After all, raw steelmaking in the UK has been an economic mess for more than 40 years. Successive governments have thrown money at the industry, nationalised, denationalised, and renationalised it without changing the central fact that the industry is unable to earn anything close to a reasonable return on capital. The outcome has been a gradual – and very expensive – contraction that has disappointed almost everyone affected as well as wasting a large amount of money.

If we were to accept that politicians and other policymakers genuinely believe their public statements, then patently they are both incompetent and have failed to learn anything from experience. The alternative is that they feel trapped and prefer to offer “noble lies” rather than trying to change the constraints that define the trap. Which is better: stupidity or dishonesty? Of course, neither is likely to burnish the reputation of the political classes.

Let us review the context. In 2024 the UK produced about 4.0 million tonnes (Mt) of crude steel, down from 7.2 Mt 5 years earlier. Total world output was nearly 1,900 Mt and even excluding China it was 876 Mt, so the UK’s output was of no significance in world terms. In contrast, Turkey produced nearly 37 Mt of crude steel in 2024, while Japan produced 84 Mt. Countries like Mexico, Spain, Egypt, Saudi Arabia and Malaysia all produced more than twice the amount of crude steel as the UK. On the consumption side, the UK’s total use of finished steel products was 9.1 Mt in 2023, while the equivalent totals were 38.1 Mt for Turkey, 12.7 Mt for Spain, and 28.5 Mt for Mexico in the same year.[1]

There are many reasons for the decline in UK steelmaking but the two most important are (a) lack of investment in new technology, and (b) poor resources. In 2023 the UK still produced nearly 80% of its crude steel by the basic oxygen process, i.e. using iron ore and coking coal in blast furnaces, whereas countries like Spain, Mexico and Turkey rely upon electric arc furnaces for more than two-thirds of their output.

Some very large producers including China, Japan, Russia and South Korea continue to rely on the basic oxygen process for more than half their output but only because they have invested in plants of size that can take full advantage of the economies of scale. Basic oxygen plants in Europe – of which there are many – are relics of a past era in the industry. New producers, especially in middle income countries, have relied upon electric arc furnaces. That includes India even though its steel producers are now choosing to build huge new basic oxygen plants, which are cheaper if they are large enough and have access to suitable resources.

An analysis of the cost structure for basic oxygen plants prepared in 2024 shows that iron ore plus coal (including transport) accounted for 64% of total costs for a notional new plant located in Japan. Hence, access to low-cost sources iron and coal are critical for anyone contemplating an investment in either a new basic oxygen plant or renovating an existing plant. For Britain the primary source of such inputs is Australia, but the country would be at a serious disadvantage relative to Japan and China in relying on resources from such a distant source.

It has been obvious for at least two decades that the only viable future for basic steelmaking in the UK has been to either (a) build a new basic oxygen plant of suitable scale on a greenfield site in a coastal location, or (b) replace existing plants with electric arc furnaces using scrap or direct reduction iron as a feedstock.[2] Both options have been resisted fiercely by steel unions and local politicians because they imply a large loss of jobs at the existing plants.

The option of building a new basic oxygen plant would raise the issue of scale and where the money would come from. The minimum efficient scale for a new plant would be an output of 8-10 Mt per year and the total cost would be £8-10 billion. The capital cost of building a new electric arc furnace plant is about one-half of a basic oxygen furnace plant but such plants have higher operating costs for feedstock and electricity.

The blast furnaces at Port Talbot have been closed and a new electric arc furnace with a capacity of 3 Mt per year is being built with an expected start date of 2028 and a reported cost of £1.25 billion. The difficulty, obvious to anyone willing to see, is that electricity prices are much higher in the UK than in other steel producing countries.

There are the usual claims that electricity from either solar or wind plants is “cheaper” than current market prices. Even if this were true (it is not!) the costs of solar and wind generation are still lower in the US and Brazil as solar and wind resources are more favourable.[4] As a consequence, the new plant at Port Talbot will only be financially viable if steel prices in the UK are higher than prices in the EU, which are already nearly double world market prices.[5]

A combination of transport costs and tariffs or other barriers to trade allow such differences to persist. Still, UK or European producers can only operate profitably within domestic or EU markets. This creates a trap because the high European price of steel penalises other producers who wish to export to third markets.

Stepping back from specific cases, the economic history of the steel industry is a classic example of industrial dynamics for large capital-intensive industries. At some point, every country wants its own steel plant – in the 1950s and 1960s in Europe, in the 1980s onwards in Asia. Lots of capacity is built and the output from new plants can be absorbed initially by rapid economic growth. However, plants age: they become less efficient, and feedstock resources become more expensive to extract. New plants have an advantage in costs and quality, so market prices decline to the point where older plants are barely able to cover their operating costs.

Inevitably, the time comes when a choice of making a large new investment to upgrade capital equipment and technology must be made. However, older plants may be poorly located or no longer have access to low-cost feedstock and other resources, so that there is little prospect that major new investments will earn an adequate return. On the other hand, policymakers and communities resist what they see as a massive loss of social capital if large industrial plants contract or close. All too often, the outcome is investment that is too little and too late – a pretence of a future that does not address the core problems.

This life cycle dynamic is not unique to the UK. Anyone familiar with the steel industry in the US and Europe can cite multiple examples over the last four decades in which the outcome has been either bankruptcy or steel plants that limp along and decline into irrelevance. All the usual reasons for sustaining basic industries have been cited in the US and elsewhere. The result is companies that are good at lobbying rather than producing steel come to dominate the sector. The current favourite justification for preserving basic oxygen furnaces is that they are necessary to produce the quality of steel required for military purposes. That may have been true in the past, but it is largely nonsense today. By using direct reduction iron in an electric arc furnace, it is possible to produce high quality steel suitable for military and other applications.

So, do policymakers never learn? After all any competent economist with experience of restructuring basic industries can provide chapter and verse for the story that I have outlined above. The problem is that there is no shortage of optimistic shills who are willing to tell policymakers what they want to hear. Restructuring is a miserable process under the best of circumstances and usually it is even worse. Few want to be known as “XXX’s butcher”, and the public reputations of too many politicians have not survived the bitter attacks from those who believe that more public money is the solution to all problems.

What incentive is there for policymakers to learn if there is no reward for following the “right” but difficult course in such cases? Politicians and other policymakers want to be liked and respected. When we talk about learning, what we really mean is accepting that the public interest may require decisions that will ensure the decision-makers are neither liked or respected by a large fraction of the public and many (most?) of their peers.

The combination of no personal accountability and the strong pressures of social conformity reinforce incentives to pretend that policies that are certain to fail have some chance of success. In the case of Scunthorpe steelworks, the UK’s House of Commons agreed without dissent to take control of the steelworks. The government will incur a loss of at least £250 million per year – and probably much higher – without any idea of how to alter that outcome. In practice, anything from £2 billion to £5 billion will be spent over the next decade without achieving anything in terms of establishing a viable UK steel industry.

Naturally, this is other people’s money. How might the electorate of Greater Lincolnshire – perhaps combined with Hull & East Yorkshire – have reacted if they had been offered the choice between spending £250 million per year on supporting the Scunthorpe steelworks or spending the same annual sum on upgrading health and other public services? Two negative responses seem plausible: (a) we don’t trust you, the money will just be clawed back in some other way, and (b) we should get the money anyway without having to make such choices.

Such views highlight why policymakers tend not to make difficult choices except under extreme pressure. They are not trusted to stick by such trade-offs – and with considerable justification. Treasury doctrine would say that money should be spent where “needs” are greatest, so the bureaucracy would undoubtedly try to unscramble any such arrangement.

The “other people’s money” philosophy infects all policymaking. No spending department ever got credit, either from the public or even from the Treasury, for being parsimonious with public money. There may be notional budget targets, but there is no real sense that public money has a high cost. One of the great failures of British governance over the last 50 years was to accept the widespread use of the term “austerity” when public expenditure was merely growing more slowly than anticipated but was not falling.

With our current tax system, the social cost of additional tax revenue (a concept that should be much better known) is between £2 and £3 for each £1 of revenue that is raised. If 50% of additional spending is wasted, which seems quite plausible, the social cost of financing an extra £1 of useful public expenditure is between £4 and £6. Can anyone convincingly demonstrate that the social benefits of spending £250 million per year to postpone the closure of the Scunthorpe steelworks exceed £1 billion per year?

To return to the original question. Ultimately, it does not matter whether policymakers do or do not learn. Policymakers are caught in a trap in which no-one believes that public money will be sensibly utilised. The primary goal is to minimise immediate criticism, even when actions will compromise any prospect of better solutions being found in the longer term. Within such a context, any expectation of a systematic evaluation of policy options, for which learning would be relevant, must be discounted.

[1] Data from the World Steel Association.

[2] Hot briquetted iron is a variant of direct reduction iron that is safer and easier to transport by sea. It is produced by reducing iron ore at high temperatures and pressures using natural gas. The EU and others are promoting the use of green hydrogen for producing direct reduction iron, but this incurs a large cost penalty, thus requiring either subsidies or tariffs to offset competition from producers less concerned about CO2 emissions.

[3] That sounds very large but JSW Steel in India claims to be spending over $110 billion (1 trillion rupees) on building a new plant in Karnataka with a capacity of 25 Mt per year, while ArcelorMittal and Nippon Steel are building a plant of similar capacity in Gujarat.

[4] The average day-ahead market price for ERCOT (Texas) was $28 per MWh for 2024. In contrast, the average day-ahead market price for the GB market was £73 ($91) in 2024. The mark-up to cover network and other costs is also much higher in the UK.

[5] The benchmark prices of hot rolled steel in late April 2025 varied from $380 per tonne in China to $750 per tonne in Western Europe and $1,020 per tonne in the US – all prices ex-works. The reference FOB price for the world market was $465 per tonne.