Post-Brexit UK economics

 “When things are bad, we take comfort in the thought that they could always get worse. And when they are, we find hope in the thought that things are so bad they have to get better.” Malcolm S Forbes

So there we have it - the British people have spoken, voting by a small majority to leave the EU. Initial reaction to the results has been largely one sided, with Remainers calling for a 2nd referendum, some Leavers “not realising their vote actually counted” (*sigh*), David Cameron resigning, and negative comments from the most esteemed policy makers across the world.

Global financial markets have whipsawed since then, likely compounding the effects of the poor performance at the turn of the year on business and consumer confidence. Though the long-term impact of Brexit will largely depend on the details of the forthcoming UK-EU agreement, a lot of the short-term impact can be attributable to recent lurches in financial markets. In this post, I step away from the much debated residual socio-economic issues and discuss the repercussions of these large financial market (equity and FX in particular) moves on the UK economy.

Equity markets

The FTSE 100 has fallen 4% since April (local highs), with the Brexit result catalysing an abrupt fall in the index. The reality is, though, that a lot of FTSE 100 companies are global companies, with the majority of their revenues coming from outside the UK (e.g metals and mining companies such as Rio Tinto, whose largest market is China). The FTSE 250, a more broad measure of actual UK companies, has taken a more severe hit with the index falling 13% in the aftermath of the referendum results. 

Though they may seem very abstract, these numbers are important representations of reality. Indeed, markets move and company market values change daily, but large falls like this can and do have an enormous impact on the UK economy in two ways. First, the effect on consumer wealth. According to the ONS, UK individuals owned 12% of the FTSE 100 in 2014 (the most recent figure). Assuming this hasn’t changed much, the FTSE 100’s fall in results day wiped out £4.7 billion off the net worth of individuals just like you and me. This is probably a lower bound as ownership would have also involved foreign stocks, and companies in the FTSE 250 which lost value too. As consumers tighten their belts following this fall in personal wealth, this tends to have an effect on consumer confidence and spending, which is no doubt negative for growth. This is a classic (and well-documented) channel by which equity markets affect consumer spending.

A more interesting channel is that of net worth. Given the asymmetric nature of information in the world we live, large loans (such as mortgages) tend to require some form of collateral for security. As equity markets fall, other assets values (such as housing) tend to fall in value too - the housebuilder stocks, such as Barratt Development, fell around 35% since the result, show how negative investors think Brexit is for house prices. With the net worth of small companies and entrepreneurs now diminished, they may not be able to put down enough collateral to obtain funding. Consequently, many of the planned expansions or investment projects will likely fall by the wayside, thus hurting (actual and potential) UK growth. 

FX markets

The cleanest way of looking at investors’ view of Brexit on the UK economy is by looking at how the GBP fared against its trading partners. It is fair to say that their view wasn’t unambiguous - sterling fell 6% against the Euro and 8% against the US on results day.

Economically (ignoring effect on the cost of my holiday this year!), a weaker Pound is more-or-less a good thing. First, the value of the UK’s exports to its markets will fall. This should fuel higher demand for UK goods as they become more competitive on the international market, via the expenditure switching effect . Importantly, this comes about absent of a manufacturer price adjustment. Moreover, the UK’s imports are now more expensive, which should urge domestic consumers to buy the same products from domestic companies. These should provide an added boost to growth. As I wrote in a previous blog post, this is by no means the end of the story - there is a whole body of work on exchange rate pass through, and it is unlikely (due to firms’ pricing-to-market and currency pricing decisions) that we will enjoy the full benefits of the weaker currency.

Second, the effect on large global UK companies cuts two ways. On the one hand, their foreign revenues and profits (if priced in local currency), will automatically rise when converted back to sterling, which could act as a counterweight to falling profits from emanating from its other markets (such as the UK). On the other hand, the financing costs of these corporates who have debts in currencies other than sterling could rise substantially. Moreover, issuance of GBP denominated debt by UK companies has more than halved since 2012. Indeed, Chart 3 in the link above shows that only around 5% of issuance by private non-financial corporations is denominated in sterling, with the majority in denominated in euros. The argument of higher funding costs for global companies, I think, carries more weight now than previously.

A lower Pound could also be positive for the current account. The current account is made up of two factors : 1) balance of trade and 2) factor income (corporate net profits repatriated from overseas subsidiaries/investments). Higher exports and smaller imports from a weaker Pound as explained above should be positive for the current account. In addition, repatriated profits that are now much higher due to currency translation should also improve the current account. 

Many city economists have revised their UK inflation forecasts up as a result from the fall in sterling. I am not so sure. As I explained with regard to the ECB and the euro, exchange rate pass through is only partial, even under long time horizons. Consequently, a fall in sterling will not lead to a commensurate rise in domestic inflation. Indeed, I would argue that though inflation will rise slightly, it will likely continue to be below the BoE’s target of 2% due to all the factors explained here. 

The bottom line

The initial economic repercussions of the referendum results will be negative as a result of increased uncertainty, lower equity markets, and a lower currency due to some of the mechanisms presented in this post. While I am sympathetic to this view, it’s important to note that the outcome could be MUCH worse if the UK did not have a flexible exchange rate policy. In this instance, it acts as a counterweight to changes in a world where prices are relatively rigid, meaning that real variables like output and unemployment (which is what we ultimately care about) do not adjust as much. In a time where the UK system is, in a sense, falling apart, it’s probably good, for our sanity at least, to bear in mind that it could be a lot worse!

The truth behind the Brexit claims: Stronger In

In the second of this two-part series on Brexit claims, I now try to provide a critical analysis of the main economic arguments of the Remain camp, ‘Stronger In’. 

1) "Over 3 million UK jobs are linked to our trade with the EU: one in every ten jobs in this country"

This statistic comes from South Bank University and National Institute for Economic and Social Research papers written in 2000 estimating that around 3 million UK jobs are directly associated with goods and services exported to the EU. The Treasury updated the numbers to incorporate 2014 data, putting the figure at 3.3 million. Note though, that the statement is NOT about the number of jobs that are at risk if the UK left the EU - clearly even with Brexit, there would still be some trade with the EU. PwC, on the other hand, estimate that 950,000 jobs are vulnerable if the UK left the EU, which arise from lower trade and investment from the EU. Perhaps this is the more pertinent question that needs to be answered. Though not incorrect, members of the public can easily mistake this and believe that 3 million UK jobs are at risk, especially in the heat of the debate. Stronger In, I think, should be more clear here, especially as this is one of their main economic arguments.

2) "Being in the EU will create 790,000 more UK jobs by 2030, creating more opportunities for you and your family"

The source of this claim is a Centre for Economic and Business Research (CEBR) paper. As has been well documented, the EU itself is not a true common market — there is not the free labour and capital flows that was envisaged when the EU was first put together, primarily due to different tax rates, laws and languages. The 790,000 figure stems from a European Parliament Added Value Unit paper which examines the benefit from further deepening of the Single Market, i.e. the benefits of the EU actually becoming a true common market. Given recent history of poor coordination between euro area countries with regard to monetary and fiscal policy, it seems unlikely that this ‘common market’ scenario will materialise in the near future. Therefore, I would argue that this 790,000 can be seen as more of an upper bound, with the more likely number being much lower. To suggest that 790,000 UK jobs WILL be created by EU membership seems, to me, somewhat intellectually dishonest and misleading.

3) "If we left [the EU], overall, your household would be worse off by £4,300 a year"

This was a statement used by Chancellor George Osborn in many speeches, and is based on the Treasury’s assessment of long term impact of EU membership. To get the £4,300, the assumption is that the UK would negotiate a bilateral trade agreement with the EU similar to the EU-Canada trade agreement. Characteristics of the agreement are: 1) no quota- and tariff-free access to the EU, and in some cases permanent exclusions for some agricultural produce and tariffs for key goods 2) continue to be outside the customs union, i.e. more administrative costs 3) reduced access because of tariffs and barriers to entry to EU markets. The report also spells out other scenarios, such as if the UK was, instead, a member of the European Economic Area (like Norway) household income would fall by £2,300. On the other hand, if the UK was just part of the WTO but did not enjoy a specific agreement with the EU, the report’s models estimate a £5,200 fall in household income. To be fair to Stronger In, £4,300 is in the middle of the extreme values. But clearly, the scenarios show just how sensitive the numbers are to the type of deal the UK strike with the EU, so a point estimate is always going to be somewhat misleading without qualification.

4) "If we left [the EU] , the cost of our imports could rise by at least £11 billion–leaving families out of pocket as prices rise"

The £11 billion figure assumes that the UK fails to reach any trade agreement with the EU, and would retain the same tariff regime currently set by the EU. Similarly to the above qualification, it’s important to note that this number massively depends on the nature of the trade agreement. A more liberal trade agreement would likely reduce this number from a direct effect point of view. Moreover, a more favourable trade deal could lead to a smaller fall in the exchange rate, therefore making this number even lower. Given the history between the UK and EU, it is likely that some deal is likely to be made, and hence I would treat this £11 billion figure as an upper bound.

5) "The benefit of being in the EU is worth £91 billion to our economy, whereas the cost is £5.7 billion"

A Confederation of British Industry (CBI) literature review proves the source of this fact, which is actually quite sizeable. The analysts think carefully about the methodology and credibility of the arguments used in each paper. In particular, they take 14 different estimates and decide that 7 of those are credible (i.e are up to date, employ a rigorous structural model of the UK economy, and have well-sourced data). Now clearly, this is where the subjective aspect of the review comes into the fore, but for our point of view, at least the process of elimination is clearly defined! 5 out of those 7 ‘credible’ estimates conclude that the benefits of EU membership outweigh the costs. Of those that do not, the assumptions seem to be ‘ambitious’ to get the desired negative result. Their review is very quick to note that no one study has captured the full effect of EU benefits, as each paper looks at something slightly different. Moreover, these differences, they argue, are complements rather than substitutes and hence the overall benefit is likely to be greater than the sum of the parts. I can empathise with this view - in something as complicated and important as this, the so-called ‘general equilibrium’ outcome is the one that matters. And things such as trade deals (on which a lot of the analysis is based) can have far-reaching consequences. The conclude that the benefits of EU membership are between £73bn and £91bn. Yet again, it seems Stronger In are using the upper bound of this in their campaign materials, so I would estimate that the numbers are somewhat lower than this.

These two posts bring to bear that both camps cite facts and research that best support their causes. This, I think, is natural in any argument, especially in one where the stakes are so high - it is, in some sense, irrational to use facts and evidence that do not support your case in any event. Though I outline and pry into the details of 10 of the main economic arguments in this debate, there are many more that need further qualification. In general, I would urge readers to take each statement with a pinch of salt, discounting any claim made by either side towards the centre. Though quantifiable evidence and numbers should, in theory, increase the transparency of ‘facts’, the reality is that the analyses themselves are very messy, and choosing the right assumptions can get you any result that you wish. Consequently, the important thing is not only the end fact or statistic, but also the assumptions used to obtain them. To paraphrase Mark Twain, we find ourselves facing another example of the indistinguishability between lies and statistics.

The truth behind the Brexit claims: Vote Leave

“There are lies, damned lies and statistics.” Mark Twain

Ever since the campaigns started, both ‘Leave’ and ‘Remain’ groups have emphasised the gravity of the decision we will have to make on 23rd June. More recently, though, there has been an urgent call for higher accuracy of evidence, to provide the public with a more complete and useful information set on which to base their vote. In this two-post series, I try to address this demand. In particular, in this post I take the main economic evidence presented on the ‘Leave’ website and aim to critique these so-called 'facts', suggest their validity, and outline what assumptions underlie the analyses in question. I do the same for the ‘Remain’ campaign in the next post.

To be clear, I choose not to impart my position on the debate in these posts, but instead take a critical view of the evidence presented from both sides.

Some of the information below is cherry-picked from a much larger House of Commons document highlighting the pros and cons of the UK’s EU membership, the rest are from various sources, cited ad hoc.

1) “The EU now costs the UK over £350 million each week”

This is the core of the Vote Leave campaign and has had its validity questioned in the public domain. Simply, all member states contribute resources to the EU, with each state paying a share in proportion to its relative gross national income (similar to economic output). So as the UK is one of the larger countries inside the EU, it naturally pays a larger share than many other countries. But what Vote Leave do not acknowledge clearly is that this £350 million is a gross figure. In fact, when taking into account what we receive in terms of agricultural, public and private sector subsidies, our net contribution is close to £190 million per week, almost half of what Vote Leave have led us to believe. Vote Leave have argued that this figure is not misleading as it still describes the money leaving the borders, so we do actually send £350million to the EU. But this again is incorrect. Contributions are calculated in arrears, so the rebate for 2014 is accounted for in the 2015 contribution, meaning that the ‘rebated money’ never leaves the UK’s borders. Leaving the EU, therefore, doesn’t mean that £350million no longer leaves the UK borders, and hence free to use for other purposes. It is worth noting too that the UK does not receive non-agricultural rebates from member countries that have joined post-2003, as per an agreement signed by the 2005 UK government. This is a point that should have been raised by Vote Leave but, to my knowledge, has not.

2) “Our EU contributions are enough to build a new, fully-staffed NHS hospital every week”

3) “Financial protection for all groups that now get money in Britain”

If the average cost of a NHS hospital is £350million, it should be clear from the discussion above that this statement no longer holds. Moreover, Vote Leave have campaigned that voting ‘Leave’ would put more money into the NHS. Note that, unlike in an election, a win for Vote Leave does not put them in power. Consequently, these savings would be put to use under the direction of the UK Government, who may not want to use them to fund the NHS. The criticism of the second statement follows naturally. Again, unlike in an election, Vote Leave would not have power (assuming the Government doesn’t change) to ensure that these groups (e.g. the agricultural sector) are indeed financially protected. In recent days, Vote Leave have put forward a manifesto for what should happen if the UK does leave the EU, but this doesn’t necessarily mean that the Government will enforce these changes were Brexit to be realised.

4) “EU regulations cost small business £600 million each week”

This comes directly from Open Europe analysis of the cost of EU regulation. In fact, this number refers to a list of top 100 most burdensome EU rules, calculated as part of the UK Government’s Impact Assessment. In reality, the costs could be much larger, which arise from administrative and implementation costs of companies caused by EU regulations. In theory, regulations are implemented with the idea of capturing or gaining some consumer net benefit. Though the costs of regulation are mostly monetary, the benefits aren’t easily quantifiable. The benefits of less asbestos in our ceilings, or higher product standards, for example, are somewhat unobservable, but it by no means suggests that they don’t exist. This is what is missing from the Vote Leave statement. Open Europe note that 95% of the benefits of regulations have not materialised, but when probed about this statement, Oliver Lewis, Director of Research at Vote Leave, said that this only applied to Energy and Climate Change regulation, and the generalisation of this was an ‘honest mistake’. Open Europe have not modified their page reflecting this…

Note that even if the UK left the EU, it would still need to abide by the standards to sell its goods and services in the EU. For example, CRD IV - the translation of Basel III regulations for the UK - will still need to be met if the UK finance sector wanted to trade with the EU. Brexit, therefore, would not rid companies of costs stemming from regulations such as these. Though the statement is somewhat accurate, the subtle implication that leaving the EU would save small businesses £600 million each week is misleading. 

Vote Leave would, and have, retorted that the spirit of this statement wasn’t that the regulations were useless, but rather that the UK should have more control of which regulations it chooses to take. I would say that this is reasonable, with the following caveat: some UK regulations are stricter than its EU counterparts, and that is a choice that the UK itself has made. That is to say the EU’s regulations are not so stringent that the UK is battling to keep up with them. Indeed, in some cases they are less stringent than those the UK imposes upon itself.

5) “After we Vote Leave, we will immediately be able to start negotiating new trade deals which could enter into force straight after the UK leaves the EU. As a member of the EU, we are forbidden from striking our own trade deals.”

For the UK to officially leave the EU, it would have to invoke Article 50 of the Lisbon Treaty. David Cameron, in a speech in Feb 2016, implied that Article 50 would be invoked immediately, and the negotiation process would start very quickly. The statement is, so far, accurate.

Sir Jon Cunliffe, Deputy Governor of the Bank of England, noted that ‘immediately’ may well not be in the UK’s best interest, as the UK first needs to know what it wants. Therefore, it might be the case that upon consultation the Government chooses to wait until it knows exactly what kind of deal with the EU it wants. Oxford Economics, Vote Leave, estimate that the average time for trade negotiations is 28 months. A paper by analysts at the CEPR take 88 regional trade agreements between 1998 and 2009 and also conclude that the average time for a trade deal is 28 months, though they are very careful to note that there is a large variance in their sample. They also mention that, unsurprisingly, the more countries that are involved in the negotiations, the longer they tend to last. Following a ‘leave’ vote, it seems reasonable to assume that a UK trade deal with the EU (which has 27 other countries) will significantly take longer than the average 28 months proposed by Vote Leave.

Though there are many more arguments put forward by Vote Leave which also require qualification, they are beyond the scope (and length!) of the blog environment. Those who are interested can find more in the House of Commons document above. These are, I believe, the 5 of the main economic arguments/'facts' that Vote Leave have used to build their case. In the next post, I address the 'Remain' camp's economic arguments in the same vein. 

Central bank solvency II: How do central banks go bust?

"We think in generalities, but we live in detail." Alfred North Whitehead

In my last post, ‘Can central banks go bust?’, I looked at the important topic of central bank solvency. I discussed the meaning of central bank solvency from a theoretical point of view and tried to dispute the common claim that solvency is irrelevant for a central bank as it can ‘print’ itself out of trouble. The common misunderstanding stems from either not knowing what ‘printing’ actually means, or from misunderstanding the role the central bank plays in money creation and inflation.  

So we now know that central banks CAN become insolvent. But HOW do they become insolvent? Traditionally, central banks has two primary functions. First, it prints actual currency (i.e. notes) if and when the economy needs it. Second, it exchanges safe, short-term government bonds for central bank reserves from designated banks in order to change short term interest rates as part of monetary policy, otherwise known as Open Market Operations (OMOs). 

Inevitably, central banks made profits from the printing of cash. The cost of printing is essentially zero, but since cash has a unique property as a medium of exchange, people are willing to give resources (which have real value) in exchange for it. This profit, known as seigniorage, has been the defining characteristic of central banks for centuries. Some estimates for seigniorage in the US range from 5% to 12% of GDP. In addition to seignorage, as they essentially converted one safe asset for another as part of OMOs, their capital position remained fairly constant. All-in-all, under old style central banking, large developed-market central banks such as the Fed, BoE and the ECB never made a loss on an annual basis, with most of the profits coming from seigniorage. 

Then came the 2008 recession, and with it a whole new style of central banking. This wave of modern central banking saw the Fed and other central banks across the world issue billions of dollars’ worth of additional reserves, meanwhile buying long-term government bonds of different risk categories, corporate bonds and, in some cases, foreign bonds. All of these carried with them their own risk: long-term government bonds carried interest-rate risk - the prices of these bonds are inherently more sensitive to changes in interest rates than short-term bonds; corporate bonds and government bonds of lower credit ratings (such as Greek bonds in the Euro area) carried increased credit default risk; and holdings of foreign bonds carried both exchange rate risk as well as credit default risk of the issuing government.

Given the complications that these unconventional monetary policies brought to central bank finances, it is clear that central bank solvency can no longer be taken for granted, and understanding it is now more important than ever before. 

Yet another question remains: why is the central bank different from any other arm of the government? Indeed, why should the solvency of the central bank be called into question when other departments’ solvency, such as the US Department of Transport or the UK Metropolitan Police Department, are not? The issue here concerns central bank independence. As has been argued for decades in academia and is now more or less regarded as fact, monetary policy is more effective under central bank independence (see my previous post on the BoE independence for a fuller review). Operationally, this can only be the case if the government has as little control of central bank finances as possible. The result, intuitively, is that the public can subsequently trust the central bank to do what is best for the economy without being shackled by government requirements and orders.

Remember though, that the government still owns its central bank. Consequently, most central bank charters outline agreements declaring how the central bank should treat profits. Typically, it entails some version of profits being remitted back to the Treasury, but how much and when are both idiosyncratic features of the central bank in question. Importantly, it also describes what happens in the event of the central bank making a loss. It is these provisions that ultimately render central bank solvency moot or not. It should, therefore be clear that the importance of central bank solvency can only be determined on a case-by-case basis by examining these charters.

The bottom line

Central bank solvency is real and an important issue. Indeed central banks do not have the Godly powers to run deficits without repercussions as some economists seem to think. Moreover, the financial crisis and the ballooning of central bank balance sheets has made this issue increasingly pertinent - in some instances, these financial barriers might even influence monetary policy decisions in smaller, less financially stable central banks. Ultimately, central bank solvency must be studied individually - a good example of thinking in generalities, but living in the detail.

Can central banks go bust?

"Some misunderstandings are hard to cure." Barton Gellman

There is a great mystique associated with central banks. Possibly because of the principal role they play in society, or perversely, because of the move towards more transparency by the world’s largest central banks, economists and the like are now more interested than ever in the workings behind them. One of the most contentious matters surrounding these enigmas is whether central banks can go insolvent. With it quite a fashionable topic since the Fed started its QE program 8 years ago, media and economists alike have given their take on this issue. In this post, I show why the common retort regarding central bank insolvency is, in my view, incorrect.

The common misunderstanding of central bank solvency is shown in the following statement: “The central bank cannot go insolvent because it can just print money.” If this sounds plausible to you, don’t worry - you’re in good company. Countless trained economists in the media and public eye believe the same thing (see here and here for some examples). If you have done any economics, you should know that, as I wrote when talking about whether the UK Government’s debt is cancellable under QE, every agent in an economy must be resource constrained. In other words: there is no such thing as a free lunch!

The first thing to understand is central bank reserves. Reserves are a form of liability for the central bank - the money that private banks leave in their accounts at the central bank must be there when they want to withdraw them. Consequently, reserves are just another form of short term borrowing for the central banks (just like a current account in any normal bank). So when people say a central bank ‘prints money to finance its insolvency’ what they mean (or should mean) is that they issue reserves to these banks and are kept in their accounts, which can then be used to ensure the central bank's solvency. Bearing this in mind, the misunderstanding in this statement becomes more clear - the central bank printing money to fund its insolvency is another way of saying the central bank is borrowing forever, otherwise known as a Ponzi scheme.

As with any Ponzi scheme, if a lender realises that something he might be investing in is a Ponzi scheme, he will not invest. In other words, the bonds that a Ponzi scheme owner issues are worthless, as no-one wants to buy them. The weird thing about a central bank is that these bonds are reserves which the central bank must (by law) exchange one-for-one for currency (cash and similar instruments). So if these ‘bonds’ are worthless, its like saying that the currency central bank issues is worthless. And what’s the equivalent of saying currency is worthless? When the aggregate price level goes to infinity - otherwise known as hyperinflation. 

There have been numerous instances of super-high inflation rates in both emerging and developed countries historically. Interestingly, most of these hyperinflation instances have come as a result of fiscal crises, where the central bank has been called upon to print money to finance the deficit. As soon as investors realise what is happening, the value of the currency that these central banks issue go to zero, resulting in hyperinflation. 

Note that hyperinflation does not stop as soon as the central bank promises to stop issuing reserves. Instead, they tend to stop with fiscal reform - i.e. when it is clear that the central bank does not NEED to print money. This is one of the reasons why intergovernmental institutions insist on fiscal reform when they are brought into countries with hyperinflation. 

The bottom line

Just to be clear, this post is not to say that one year of the Federal Reserve, for example, printing money to cover its shortfall is going to lead to hyperinflation in the US. There is a large amount of credibility that the Fed has built up over many, many decades that will prevent this from happening (not to mention the laws and public scrutiny that it will come under). But I trust that this threat to the Fed’s existence is enough of a deterrent to prevent the Federal Reserve Governors to ever think about doing such a thing. 

This post does, though, show that from an economic perspective, in the long term (i.e. in equilibrium) a central bank CANNOT print money to keep itself solvent. And in doing so, it can indeed go insolvent. Though this may be said many, many times in all kinds of public forums, the common belief that central banks can print itself out of insolvency will likely continue. Indeed, some misunderstandings are hard to cure.

Are expectations rational?

“You can't base your life on other people's expectations.” Stevie Wonder

Ever since the revolution of the 70s and 80s, rational expectations, popularised by Robert Lucas in 1972 have played a fundamental role in macroeconomic theory. Before this, it was assumed that agents' expectations were adaptive. Essentially, they were backward-looking to the extent that their expectations of any variable were merely last period’s actual reading. 

Lucas and the revolutionists, however, suggested that people were much cleverer than this. In particular, he declared agents forward-looking, with their expectations of the future developed by taking into account past and present events in the correct way, naming 'them 'rational'. The beauty stems from it (broadly) making intuitive sense. Simply, if you think that the price of a car will be 10% higher tomorrow, you're likely to buy it today instead. That is, your actions today reflect your expectations of what will happen in the future.

Formalising rationalising expectations

The use of math in economic theory is both a blessing and a curse. On the plus side, it means that there’s no room for ambiguity - there’s no ‘I said x, but what I meant to say was y’ argument to be had, because math requires precise definitions of every variable. But the problem is that sometimes you have to define things that are really hard to define. One of these is rational expectations. 

So how were they defined? For a long time, rational expectations assumed agents to have perfect foresight - their expectations of inflation in a model were exactly what inflation came out to be. While this might seem ridiculous at face value, slightly deeper reflection leads you to the idea that rational expectations means that, at least in the long run, expectations of, say inflation, coincide with actual inflation outcomes. This, on the other hand, does not seem as strong an assumption. 

Even so, how does this theory stack up against the evidence? Is it another one of those implausible assumptions that the economics profession makes? In the remainder of this post, I pit the theory against the data using three basic tests of rationality

The data comes from the Survey of Professional Forecasters (SPF) for the euro area - a quarterly survey of expectations for some of the euro area’s key macroeconomic variables. Importantly, the panel consists of around 90 financial sector and non-financial sector forecasters from a range of different countries within the EU. The choice of this survey compared to other surveys ties into the belief that rationality itself is quite hard to achieve. As such, if anyone can satisfying the theory, people who are paid to forecast for a living should have the best chance of doing so. The tests and analysis I conduct stems from Mankiw and Reis (2004), who conduct a more elaborate version of this study using US data and surveys from multiple sources.

The Data

These tests of rationality are conducted using the data of HICP inflation expectations (source: SPF) and actual in the euro area from1999 to Q4 2015 (source: Eurostat). 

First a word on the data: looking at the period as a whole, forecasts seem to have been always too optimistic. As you can see from the histogram above, forecast errors (defined as actual inflation minus expected) take readings from -0.9 to -2.5, entered around -1.5. This seems to be quite stable, as splitting the period up into pre- and post-crisis doesn’t yield significantly different distributions. Indeed, the data suggest that forecasters have always been disappointed when it comes to inflation in the euro area.

1) Bias

The first test of rationality checks whether, on average, forecast errors are unbiased. That is, forecasters should not have preconceived notions on how inflation will turn out that affects their forecasts. In a deeper sense, the theory claims that they take into account previous data that leads their expectations to be, on average, correct. In terms of the data, this would be represented by a regression of the forecast error against a constant. Given we would expect rational agents to be unbiased, we would expect the coefficient on the constant to be 0. 

We can see the results of this regression in column 1 of Table 1 below. The coefficient is negative and significantly different from zero at the 1% level. Therefore, the data suggest that economist forecasts are not unbiased, and specifically biased upwards. It’s worth noting that result still holds when we split the data up in to pre- and post-crisis periods.

Conclusion: Test 1 refutes the rational expectations theory claim.

2) Persistence

Here we check whether forecast errors are correlated with those in the past (specifically, those four quarters ago - in line with Mankiw and Reis (2004)). Rational expectations theory tells us that if agents notice that their errors were, say, too low in a previous forecast, they should internalise their error and correct for it in the future. Empirically, we look at a regression of errors against its own lag. This would take the form of the coefficient on the lagged term equal to zero if the rational expectations hypothesis were to hold. 

As we can see in the second column of the table, the coefficient is not different from zero, even at the 10% significance level. But note that the coefficient is significantly different from zero when we look at the pre-crisis period (not shown here). Though we choose to discount this given this particular regression has only few (33) data points, it is still noteworthy.

Conclusion: Test 2 confirms the rational expectations theory claim.

3) Forecast revisions

One can only form expectations given the current information set available. But as new data come in, rational expectations would require that the new data is fully incorporated into the new forecast. As such, for rational expectations to hold, it must be that the change in the forecast (revisions) should not tell us anything about the forecast error in period t.  As such, a regression of the forecast error on forecast revisions should gives us a slope coefficient of 0. 

This is shown in column 3 of the table. The slope coefficient on revisions is not significantly different from 0 (even at the 10% level), implying that these forecasters do fully (in the statistical sense) incorporate new information into their revised forecasts. It is, again, worth noting that this result holds in both the pre- and post-crisis periods. 

Conclusion: Test 3 confirms the rational expectations theory claim.

Table 1

	Forecast error		Forecast error
	(1)	(2)	(3)
Constant	-1.52***	-1.08***	-1.53***
	(0.08)	(0.32)	(0.06)
Forecast error lag (t-4)		0.30
		(0.22)
Forecast revisions (t-1)			0.48
			(0.32)
Observations	64	61	61
R2	0.00	0.10	0.05
Adjusted R2	0.00	0.08	0.03
Notes:	***Significant at the 1 percent level.
	**Significant at the 5 percent level.
	*Significant at the 10 percent level.
	Standard errors in parentheses; Newey-West standard errors to correct for serially correlated errors, lag = 4

The bottom line

Rational expectations have been one of the most influential concepts in macroeconomics, playing a part in almost all mainstream economic theory since the mid-70s. Moreover, it’s feature of forward-looking individuals has given rise to unconventional policy proposals such as quantitative easing and forward guidance, that arguably prevented the world economy from falling into another Great Depression. But how realistic is it? Through a euro area inflation expectations lens, the data suggest that EU inflation forecasters exhibit something close to expectations that are rational. Mankiw and Reis (2004) also find something that broadly resembles rationality when using the US SPF. Indeed, and somewhat contrary to popular opinion, the rational expectations tenet is not something far from reality. 

Stevie Wonder once said ‘you can't base your life on other people’s expectations’. While that may be good life advice, let’s just hope, for the sake of our economic models, that you at least base your life on your own!

Oil price effects: a simple evaluation

After some time off for exams, the blog posts can now continue! I'd first of all like to wish you all a happy and prosperous New Year. Now with the formalities over, let's get on with it!

"Everything changes but change." Israel Zangwill

Unless you’ve been living under a rock for the past one and a half years, you’d be well aware that the price of oil has been having a torrid time. From a local peak of $115 per barrel in the summer of 2014, Brent crude oil prices have fallen to just under $27 a barrel - a fall of around 75%, even more than its fall doing the worst of the 2008 financial crisis. Over the past year, the macroeconomic effect of this has been one of the main areas of debate, with many respectable economists and institutions giving their two cents. Now that enough time has passed to gather some actual data, this post takes a step back and tries to figure out who was right and wrong, and more importantly, what the actual outcome was.

There were two schools of thought regarding the reasons behind the oil price weakness at the time - one based on the supply side, and one on the demand. To be clear, I stand somewhere in the middle - I believe both sides have played a role, but the timings of the price fall suggest it is more a supply-side issue than that of demand. 

The supply-siders:

The hypothesis here was that the oil price weakness was due to a large increase in oil supply (see the ECB here). Here, charts were shown of the rapid expansion of shale oil produced in the US which flooded the market. At the same time, quotes and stylised facts were reeled off regarding the decision by OPEC (aka Saudi Arabia) not to cut production in the wake of oil price weakness, unlike previous actions at similar junctures. This, according to them, signalled to markets that the ‘reaction function’ of the Saudis had changed. I have some sympathy with this view, and it has not be more present than the lurch lower in the oil price after the lifting of the Iran sanctions that should bring 500,000 barrels a day to the market. 

Subsequently, as the world was hit by a supply shock, proponents of this hypothesis concluded that input costs of net oil importing countries (mainly advanced economies such as the Eurozone, UK and the US) would also fall, allowing for - at the margin - higher consumption and production and hence economic growth. (See the IMF’s take on the issue here.) On the other side, those countries who were producing this oil would for sure feel the brunt, leading to weaker growth and a depleting of their fiscal revenues. But as these countries were small in GDP terms, considering the problem from a global perspective would leave world in a much better place.

The demand-siders:

The other school of thought posited that the oil price weakness was a symptom of already weakening global demand (here and here). It was pointed out that the oil price has already started falling before the OPEC announcement and that countries across the world were already rife with disinflation. Furthermore, a slowing China was the smoking gun that brought the argument together for this group. 

Weak global demand, they said, was predominantly due to the ‘debt overhang’ from the financial crisis. That is, households and firms, who took on too much debt when times were good, were now having to cut back on investment and consumption to pay that back. So, when faced with such a windfall from this lower oil price, agents would either save it or use it to pay off debt - exactly what caused the weak demand in the first place. Consequently, consumption and production wouldn’t rise, meaning that we wouldn’t see the increase in GDP that the supply-siders were expecting. 

What does the data say?

Unfortunately, the answer isn’t so clear cut. The world economic output grew by 2.4% from Q1 2013-Q1 2014, just before the oil collapse. The same period in 2014-2015, GDP growth fell to 1.72%. Some commentators on the demand side have used this as confirmation of their theory. But this link is tenuous at best. With global GDP consisting of a multitude of factors, not just consumption, evaluating their modelling of global consumption with respect to a change in the oil price this way could be misleading.

The chart above shows private consumption of the 5 largest net importers of oil. The idea here is that these countries are supposed to have benefitted the most from cheaper oil as they are the ones that use it the most. Note that these 5 countries account for 55% of the world’s GDP, a large chunk. As you can see, nearly all the countries have seen positive changes to their private consumption growth compared to a year previously. Those who do not (the US and China) are marginally negative. So a closer look suggests that, cheaper oil HAS been positive for consumption in these countries. 

So a slight dip into the detail, it seems that those arguing for the increase in consumption in developing countries were right. It turns out that consumers did go out and spend more than they did in the previous year - which I attribute to the fall in oil prices. On the other hand, looking at the aggregate data, it would seem that the demand-siders were right - world GDP and indeed GDP growth in a lot of those oil importing countries didn’t rise substantially.

The bottom line

This dichotomy, I think, really boils down to a subtle assumption in a lot of economic models when trying to uncover effects such as these. The assumption is always ‘ceteris paribus’, or ‘all else equal’ (in less fancy language). This is amounts to what economists call partial equilibrium analysis. The question they are essentially asking is, ‘if oil prices fell, and nothing else changed, what would be the effect of output and inflation?’ 

The reason why partial equilibrium analysis is so common is that it simplifies the problem into bitesize, easily digestible, chunks. Though this may be useful to understand the causes of things, it may not correctly forecast the actual outcome, because things that were assumed to be unchanged do change. This is where DSGE models come into their own, as they aim to forecast the direct and indirect effects of a change in a variable - in this case, the oil price. Nonetheless, they assume ‘exogenous’ factors outside the model are constant. But again, even when evaluating these models, it’s worth bearing in mind that the only thing that doesn’t change is change itself. 

The truth behind US growth

“Let us make future generations remember us as proud ancestors just as, today, we remember our forefathers.” Roh Moo-hyun

For a lot of macroeconomists, the 2008 financial crisis provided the much needed kick-up-the-backside that was missing throughout the previous two decades. From their ivory towers, they reflected on what had been missed, what should have been done and, more pressingly, what could be done to resolve the situation at hand. On the ground though, the crisis was whole lot more than just another intellectual game. Jobs were lost, lifetime savings were wiped out and millions of lives were ruined. This post tries to explain US economic growth since the crisis, and forms an argument for easy government policy for some time in the future.

The chart above shows US real GDP per capita (GDP adjusted for population and inflation) from 1860 to 2000, taken from Jones (AER, 2002). It shows that, incredibly, living standards in the US (proxied by GDP per capita) has grown more or less by 2% per year for the past 150 years. There are two obvious blips - the major contraction of GDP in the 30s, called the Great Depression, and the following boom, known as World War 2. Updating this chart to today, we get the chart below, which shows US GDP growth per capita from 1960 to 2014 (look at the red line, for now). The yearly growth in GDP per capita in this period also sits around 2%. Amazingly, almost as if a law of nature, the US has grown (in per capita terms) by around 2% per year since 1860! What are the driving forces behind this incredible feat?

The theory

When thinking about growth, it is hard not to start with the Solow model. The Solow model, first characterised by Robert Solow in his seminal 1956 paper, characterised economic growth per capita as dependent on two things - technological progress and capital accumulation. In particular growth, it states, can occur in two ways:

1) Through the accumulation of savings, which is subsequently invested into new capital (infrastructure, buildings, know-how etc). 

2) Through technological change, which expands the frontier of growth possibilities for an economy (for example computers, inventions, innovation etc). 

The problem with option 1 is that there is an implicit ceiling on the proportion of income you can save - 100% (assuming no borrowing). Consequently, there is an implicit cap on how much an economy can grow. During the 1960s, Russia chose this growth option and found out the hard way. By reaching the maximum proportion they could feasibly save, they saw their growth and economic status fall by the wayside. The important implication of the theory therefore, is that for a country to go infinitely, it has to be primarily through the channel of technological change. This seems to be the case with the US over the past 155 years. Indeed, the savings rate has remained fairly constant through this time too, further suggesting that growth was mainly through technological change. 

The common misdiagnosis of growth

When referring to the financial crisis, commentators tend to link the downturn with a fall in the capital stock. I hear phrases like ‘the financial crisis destroyed some of the capital stock’ or ‘the US capital stock was cut to below equilibrium levels during the financial crisis’ over and over again. Once the capital stock falls, so does its depreciation. And so, the model dictates, as savings remain unchanged, accumulation of capital (=savings) is above the rate of depreciation causing the capital stock to rise until these two  are once again equal - in equilibrium. This means that GDP rises to the same level it was before the crisis and then continues growth from there. This would look something like the blue line in the chart on the right. Unsurprisingly then, these same commentators end up confounded when GDP looks like the red line instead - there hasn’t been this ‘catch-up’ growth, but rather the economy just grows at the same rate but from a lower level.

What if the crisis has affected the US in the second way? - a reduction in technological change, otherwise known as a negative productivity shock. Put simply, the financial crisis has resulted in a onetime reduction in the frontier of economic growth, pushing GDP lower. But importantly, there is no ‘catch-up’ because the equilibrium level of GDP itself has fallen. This is actually quite plausible. Given that investment these days is predominantly funded by financial markets, its distortions following the crisis could easily have made funding for innovation and technological advancements more difficult. When added to the increased risk-aversion by companies and hysteresis effects of long periods of unemployment, it is quite conceivable that the crisis resulted in a negative productivity shock to the economy.

One problem with the Solow model is that it doesn’t specify how technological progress happens - it is assumed, in economic parlance ‘exogenous’, some sort of manna from heaven. To understand how innovation and technological change occurs, we look towards other theory, namely Endogenous Growth Theory. The specifics of the different theories are irrelevant for our purpose, but the important takeaway is that for innovation grow constantly, it must be linear in itself - i.e. it cannot depend on the level of innovation. No matter what level of innovation there is, the growth in innovation will restart itself. So for a country already at equilibrium and growing predominantly through technological change, growth restarts itself at any point. Indeed this is exactly what we see in the chart - there was a one time reduction in GDP, but then its growth continued from the lower level (the red line). 

The bottom line

Ultimately, policy makers should aim to reduce the reduction in welfare stemming from such crises. If the recession was due to the destruction of capital, the welfare loss would be small - something like the (badly drawn!) blue shaded area. But if it is due to a negative productivity shock, as this post suggests, the welfare loss is very large - the red area plus the blue area. Note the red area will continue to grow as time passes. Policy makers should try their utmost to bring the economy as quickly as possible to the blue line, with easy monetary policy and increased government spending. 

When faced with a crisis, policy makers should do all in their power to achieve two things. First, to make the current generation’s lives as prosperous as possible given the current circumstances. And second, ensure future generations do not pay for our mistakes today. By violating these two tenets of policy making, we may well find ourselves making the lives of future generations not better, but worse.