Enemy of the steak: what's wrong with government diet guidelines

As an amateur chef I have become increasingly interested in the government’s guidelines and regulations around food. For something so central to our lives, the advice and rules the government makes to do with what we eat are usually overlooked. Two developments this week suggest that this is a mistake. I have previously argued that government regulation is often bad because, if it turns out to be bad regulation, it imposes a single error across an entire group of people or firms. That view may explain the financial crisis, where banks were required to hold lots of mortgage debt by regulators who thought they were forcing banks to be sensible.

Now, it looks as if it might also apply to diet guidelines. This week a new paper has been published that argues quite convincingly that, not only does modern evidence show that government guidelines to reduce dietary fat intake were a bad idea, they were even against the bulk of the evidence available at the time.

Today, it’s being reported that the US will stop advising people to avoid dietary cholesterol, because of a change in nutritionists’ view of how our diet affects our bodily cholesterol levels.

The Verge says that ‘The DGAC is more concerned about the chronic under-consumption of good nutrients, noting that Vitamin D, Vitamin E, potassium, calcium, and fiber are under-consumed across the entire US population.’ Interestingly, high-cholesterol foods like eggs, offal and seafood are very high in some of those vitamins.

It’s tempting to suggest a connection there – that vitamin deficiencies may be a direct cause of misguided government diet advice. And this may be the case. But, having looked around and spoken to the British Nutrition Foundation, I can’t find any work by either the government or independent academics on how much impact these guidelines have on what we eat, let alone on our health. (The exception is the five-a-day campaign, which has been fairly successful.)

If it turns out that diet guidelines have been wrong on things like fat and cholesterol, and maybe things like salt as well, what are the costs? I see there being two potential downsides to bad advice. The first is that the advice is actually dead wrong and drives people to eat in ways that ends up being worse for their health. Perhaps this is true of the cholesterol advice.

The second, which is more ambiguous, is the welfare cost. We eat not just for sustenance but because it gives us pleasure – a steak done well is much better for me than a well-done steak, because, even though the nutritional content is basically the same, it makes me happier. If government guidelines have been mistakenly putting people off eating foods they enjoy then they have been costly in welfare terms even if the health impact is not significant.

Of course people may need to get advice from somewhere, and I don’t see any reason to believe that government advice is worse than, say, the stuff you get in the Femail section of the Mail Online website. But if government diet regulations are still likely to be mistaken, and they influence people much more than any single bit of diet advice from an independent source, then they may end up holding back a process of private trial and error that would give us better information about what’s good to eat over time.

Where the US justice system is and isn't racially biased

In a timely post Scott Alexander investigates the evidence around the US justice system to see where, if at all, it is systematically biased against African-Americans. He looks at quite a lot of empirical evidence and concludes that:

There seems to be a strong racial bias in capital punishment and a moderate racial bias in sentence length and decision to jail.

There is ambiguity over the level of racial bias, depending on whose studies you want to believe and how strictly you define “racial bias”, in police stops, police shootings in certain jurisdictions, and arrests for minor drug offenses.

There seems to be little or no racial bias in arrests for serious violent crime, police shootings in most jurisdictions, prosecutions, or convictions.

This is important given the news coverage of the killing of Michael Brown, an 18-year old African-American in Ferguson, MI, by a white police officer. Although a lengthy grand jury investigation found that the police officer did not act unlawfully, many have rejected this verdict.

They may be motivated by a belief that the justice system is predisposed to exonerate white police officers who act wrongfully to racial minorities. This is a phenomenon I have written about in the past – one has to use one's existing beliefs to assess new information to make sense of the world at all. But Alexander's investigation of the evidence suggests that things may not be as clearly biased as they believe (or, indeed, as I did before reading the post myself).

Alexander makes an important point, however. Although law enforcement may be less biased in the US than we think, the laws themselves may still be very biased (even if that bias is unintended, which perhaps it is). Drug laws, which seem extremely unjust, will cause more injustice to African-Americans if they use drugs more regularly than other racial groups. And then there is the fact that African-Americans may be poorer on average than than white Americans, so they cannot access the same quality of legal defence.

The lesson from this may be that, though we can never escape the 'webs of belief' we construct to understand the world, we can try to be aware of the fact that we use these. If instead we decide to view disagreements about politics as existing because bad guys have incentives to fight good guys, we may end up in dark places where no amount of evidence will ever convince us that we may be mistaken.

It ain't what you don't know that gets you into trouble

It's what you know for sure than ain't so that does. And Mark Twain's observation has a great deal to tell us about how and why economic policy is so often so bad. For it's not just the politicians who are deeply misguided as to what the facts of the situation are: we the citizenry can be alarmingly inaccurate in the things we believe about the economy. As Tim Taylor points out:

There's an old saying often attributed to Daniel Patrick Moynihan that "Everyone is entitled to their own opinions, but not to their own facts." In public opinion surveys, of course, people are offered a chance to assert facts that reflect their own frame of mind. For example, Social Security is popular, while foreign aid is not, and therefore people (wishfully) hold the opinion that we must not be spending too much on Social Security, but are spending a lot on foreign aid that could cut with little domestic pain. But it's obviously tricky to have a productive social discussion about economic issues when there is little agreement on central facts.

Very much the same holds true in the UK: people overestimate the unemployment rate, how much is spent on overseas aid, underestimate how much pensions cost and so on. And then of course there's the very slightly more tricky things that we should be able to agree upon but generally don't: a higher minimum wage reduces the number of jobs, rent control is the best way of destroying urban housing short of aerial bombardment and so on. We even have our own version of that Moynihan quote, comment is free but facts are sacred.

Would that public discourse, the setting of public policy, took place within the boundaries of those facts rather than being misinformed by what people are sure is true but just ain't so.

Voters are very ignorant, and that should terrify you

Voters are very ignorant about the basic facts of politics. This is where Americans fall when asked what the US government spends the most on: And here is how the money is actually spent:

As I've often asked before, how can we possibly expect voters to elect the right people if they know so little about the issues at stake? It's like asking a blind man to be your ship's navigator.

Governments have vast powers and responsibilities. Their reach is essentially limitless. And the people who decide what they do are hopelessly ill-informed about the world. Forget the Hayekian knowledge problem – the voter ignorance problem means democracies cannot hope to elect decent governments with the priorities and policies that the voters themselves would want if they were well-informed.

Elite rule might have been the answer, but elites are dogmatic, closed-minded ideologues. No, there does not seem to be any group we can rely on to rule. Voter ignorance should make us extremely reluctant to bring the state in to solve some problem we're having.

And before you tell me that democracy is the worst system we know of, apart from all the others: Are you sure?

Two cheers for technocracy

Who needs experts? The minimum wage was once an example of the triumph of technocracy, where decisions are delegated to experts to depoliticise them. The Low Pay Commission was set up to balance competing priorities – increasing wages without creating too much unemployment. If you were a moderate who thought the minimum wage was a good way of boosting low wages, but recognised that it might also create unemployment, the LPC gave you a middle ground position. (For what it’s worth, I’m an extremist.)

That technocratic settlement also allowed politicians to, basically, safeguard against an ignorant public. By delegating decisions like this to experts, bad but politically popular policies could be avoided. Relatively well-informed politicians could avoid having to propose bad policies by depoliticising them.

Other examples of this include NICE’s responsibility for deciding which drugs the NHS should and shouldn’t provide, and the Browne Review that recommended student fees, which had cross-bench support. The old idea that “you can’t talk about immigration” comes from an informal version of this – everyone in power knew that people’s fears about the economics of immigration were bogus, so they were basically ignored.

But that technocratic settlement now looks dead. Labour has now made a specified increase to the minimum wage part of its electoral platform, following George Osborne’s lead earlier this year. That means that voters will have to choose not just between two rival theories about the minimum wage, but two competing sets of evidence about whether £7/hour or £8/hour is better, given a wage/unemployment trade-off.

Whether voters are self-interested or altruistic doesn’t really matter. A self-interested low wage worker would still need to know if a minimum wage increase would threaten her job; an altruistic voter would similarly need to know a lot about the economics of the minimum wage and the UK’s labour market to make a judgement about what level it should be.

And of course the minimum wage is just one of dozens, if not hundreds, of questions that political parties offer different answers to that voters have to make a judgement about.

In practice this does not happen. Voters are very uninformed about basic facts of politics, and are almost entirely ignorant about economics, which almost everyone would agree would be necessary to make the correct judgement about something like what the minimum wage level should be (even if they didn’t agree on which theories and evidence was relevant). Even the use of rules-of-thumb such as listening to a particular newspaper or think tank (ha) will suffer from the same problems.

Voters, then, face a nearly impossible task. Assuming they are bright, well-intentioned, and believed that it was important for them to cast their vote for the party that would have the best policies, they would have to amass an enormous amount of information to make the right decision on all the questions they, in voting, have to answer.

So voters are trapped. They cannot know what minimum wage rate is best any more than they can know what drugs the NHS should pay for. They are, empirically, very unaware of basic facts, but they would find it hard to overcome that even if they wanted to.

Does democracy make us free? Maybe, but it’s the freedom of a deaf-blind man – we can choose whatever policy we want, without any idea about what those policies will actually do. So, if the alternative is more direct democracy like this, maybe technocracy isn’t so bad.

When ignorance trumps incentives

When something bad happens it is often helpful to think about why it has happened in two ways: did someone have a reason to make it happen, or did it happen by accident? This can also be expressed in a slightly different way: were incentives to blame, or ignorance? Jeffrey Friedman and Wladimir Kraus have made a compelling argument that ignorance explains more about the world than we often realize, using the 2008 financial crisis as an example. This post is an attempt to summarise their argument.

Economists often remind us that incentives matter. Indeed this is sometimes said to be the cornerstone of ‘the economic way of thinking’. Russ Roberts gives the example of death rates on British ships bringing convicts to Australia in the 18th Century – rather than attempting to raise ship captains’ awareness of the badness of letting their passengers die, the government gave captains a bonus for every convict that walked off their ship. This was very effective.

Clearly this way of thinking can be very powerful. It is the foundation of the price system, which is the mechanism that markets use to allocate resources effectively in a world where information is dispersed: if demand for pizza rises, the price of pizza rises, giving cooks and restaurant owners an incentive to sell more pizza. It helps to explain why some people stay on welfare payments for long periods of time: the welfare money they lose when they go into work represents a significant disincentive to work. Or, if you offer something like a bailout to businesses that go bust, you reduce the incentive for them to act prudently to avoid going bust.

This last example is what is known as moral hazard. And it is a popular and compelling explanation for the 2008 financial crisis. Banks expected to be bailed out if they went bust, so they acted more recklessly than they would if they thought they would be on the line for their mistakes.

However, Friedman and Kraus argue that this popular and compelling explanation may in fact be wrong. A good way of testing it would be to compare how the bankers involved in making bad decisions acted where perverse incentives applied, and how they acted where perverse incentives did not apply.

One strong piece of evidence against the incentives narrative is that bankers seem to have acted the same way with their personal investments as they did with their business investments.

Many bankers lost a lot of money personally in the crisis because their personal portfolios were not ‘bailed out’ in the same way that their banks were. If we are to treat the ‘incentives story’ as a falsifiable proposition (as all claims about the world should be treated), this might be a fairly strong reason to disregard it.

This may be where ignorance comes in. If bankers acted the way they did because they were unaware of the risks they were taking, then we would expect their private and business investments to be pretty similar.

However, it is strange that so many bankers seemed to make the same mistake. We know that they were not acting in a neutral environment: as Friedman and Kraus have shown, regulations like the Basel accords and the US’s recourse rule directed banks to prefer mortgage debt to business debt. Other regulations directed banks to rely on the risk judgments of three specific ratings agencies, giving those agencies protection from competition.

(On the ratings agencies point, astonishingly, it seems that nobody realized that these agencies were basically protected from competition. Both bankers and regulators assumed they were being subjected to market forces, leading to everyone trusting them a lot more than they would if they knew they were dealing with protected monopolies.)

These regulations were designed to make banks act prudently: the regulators had no incentive to make banks act badly. It seems possible that they did not realize the error of their ways until it was too late. Perhaps regulatory ignorance was to blame.

It is important to stress that the regulators should not be blamed personally. They probably made the best choice they could have made given the information available to them. Rather it is the position they found themselves in that seems to have been to blame. If a single bank (or even a handful) makes a mistake, that bank will suffer but the whole sector probably won’t. It is only when a whole sector of a market (or almost all that market) makes an error that we should worry. (Incidentally, as shaky as the housing and financial sectors were, the real trouble did not begin until monetary policy tightened unexpectedly, as Scott Sumner outlined at our recent Adam Smith Lecture.)

Given ignorance, we should expect errors to take place. Because regulation necessarily applies to everyone in a market, a regulatory error affects everyone.  That may be the fundamental problem with regulation, and a reason to have a strong ‘prima facie’ objection to regulation. It is better to have one hundred firms making one hundred different mistakes that happen at different times and in different ways to one hundred firms making one single mistake that happens at the same time for everyone.

None of this implies any special knowledge on the part of firms. Indeed regulators may be much more expert than the firms they are regulating, but the danger of a collective error would still give us a reason to generally object to regulation in principle, no matter how sensible it may seem.