24 August 2017

When will we learn to take economic predictions with a pinch of salt?

By

So, it turns out that the UK’s current account deficit, already the biggest of any major economy, is even bigger than we realised. After a series of revisions, the ONS figures now show that Britain borrowed almost £100 billion from international creditors in 2015, £18 billion more than earlier estimates. As a result, the current account to GDP ratio for 2015 is actually 5.3 per cent rather than the previous estimate of 4.3 per cent.

Leaving aside the question of whether we should be worried about the current state of the balance of payments, the debacle just goes to show how difficult it is to compile accurate economic statistics.

Almost all economic statistics are based on survey data. Take inflation for example. This is measured by a team of people in different parts of the country surveying various things such as prices in supermarkets. As for GDP, forms are completed by various companies and are sent back to the ONS. These figures then go into complicated mathematical models and out pops GDP.

These numbers are frequently revised up and down, sometimes resulting in recessions being erased as was the case in 2013. The coalition government received a slew of negative headlines over the apparent double-dip recession and calls from the opposition to change policy. But the idea that policy should be dramatically changed because an estimate of something incredibly difficult to estimate has moved slightly in one direction or the other is patently ridiculous.

Sir John Cowperthwaite, the financial secretary of Hong Kong during the 1960s, famously refused to collect all but the most perfunctory statistics, arguing that they would only end up being used to justify intervention and excessive economic planning to correct apparent problems. If these statistics turn out to be inaccurate, the potential problems are amplified by ill-informed policy responses.

Would George Osborne have introduced a Lifetime ISA and dramatically increased ISA allowances had he known the savings ratio was significantly higher than the ONS initially estimated for the past decade?

The problem is even worse when it comes to forecasts.

Instead of being viewed for what they are – namely, predictions – forecasts are regarded by the government and large swathes of the media as facts. The most infamous example in recent history of the failings of forecasting was during the Brexit referendum in 2016.

In the build up to the vote, HM Treasury and the Bank of England released forecasts based on their models which predicted doom and gloom for the economy if the British people voted to leave the European Union. When economic meltdown did not befall the United Kingdom in the days, weeks, and months following the referendum, questions were raised as to the validity of these forecasts. However, a year on, we seem to have forgotten the lesson of this episode, with predictions made by economists still regarded as gospel by politicians and the media.

The example of the forecasts produced by the Treasury and the Bank of England during the referendum serve as a reminder of the importance of economic forecasts. Many people were no doubt influenced by these forecasts from respected institutions which are regarded as facts and so this would have had an impact on how they decided to vote. However, not only are they influential in the democratic process, forecasts are also relied upon to decide government spending and policy decisions. These decision can have serious consequences for the health of the economy and also have huge impacts on the lives of the citizens of that country.

Many economists have admitted that the reputation of their profession was severely damaged by the failure to predict the financial crisis. For example, Andy Haldane, the Bank of England’s Chief Economist called it their “Michael Fish Moment” in reference to the weatherman’s failure to realise that a hurricane was going to hit the UK in 1987. Haldane pointed out that this led to meteorologists developing more sophisticated models which produced more accurate forecasts. This, he assured the Treasury Select Committee, was now the case with economists.

However, there are real problems with official statistics and economic forecasts due to how they are formulated.

The second issue is that they are based on mathematical models. These models make various assumptions such as perfect information and perfect competition. Moreover, they take the point where supply meets demand and calls that the equilibrium price. However, perfect information, perfect competition, and the equilibrium price do not exist in reality. When models have such concepts built in they have assumed away the real world.

It is a desire to treat economics as a science with formulae and equations which results in forecasts being so frequently wrong. Economics should not be regarded as a science such as physics or chemistry where the movement of objects and the reaction of molecules can be predicted accurately as they obey the immutable laws of nature.

The economy is not like this because people are not like this. As the famous Wall Street Journal journalist, Henry Hazlitt once wrote: “The economic future, like the political future, will be determined by future human behaviour and decisions. That is why it is uncertain. And in spite of the enormous and constantly growing literature on business cycles, business forecasting will never, any more than opinion polls, become an exact science.”

People are not like atoms or planets. Our behaviour is not governed by universal laws. We are unpredictable. Moreover, events are unpredictable and so it is incredibly difficult to say what will happen in the economy in the future as we are not psychic- we do not know what will happen in the future and we don’t know how people will react. In short, it is extremely difficult to predict the unpredictable.

Furthermore, looking to the past is also not a surefire way to predict the economic future. Just because people acted in a certain way to various events in the past is no guarantee that people will react in the same way to similar events in the future. As the Austrian economist Ludwig von Mises pointed out, “external phenomena affect different people in different ways” and “the reaction of the same people to the same external events vary”.

This is the fundamental problem facing the economics profession, the reliance on mathematical models which are based on assumptions. The great thinker John Stuart Mill argued that nobody could be a good economist if they only had an understanding of economics; he believed that they also needed a good grasp of philosophy, history, and the human condition. This was true in Mill’s day and it certainly true in ours.

To understand the economy we cannot simply rely on assumptions and equations, we need to understand what motivates people to act as they do. To achieve this, economists need to grapple with disciplines such as psychology, philosophy, history, and sociology as well as mathematics.

Economic models do serve a useful purpose. However, they should be treated with caution and a healthy degree of scepticism. They are not quite akin to gazing into a crystal ball or reading tea leaves, but they are not a fool proof method of predicting the economic future.

Ben Ramanauskas is a Policy Analyst at the Taxpayers' Alliance