My last piece “Fiscal Charter: it works BECAUSE it’s a gimmick” attracted some interesting responses on twitter. Well regarded academic and former Bank of England committee member Danny Blanchflower was kind enough to comment, despite my rather forthright comments about him in the piece.
Critics made a couple of points:
- Carmen Reinhardt and Ken Rogoff (whose data I referred to when demonstrating that governments are serial defaulters and that debt rises for cyclical reasons after a banking crisis) had made errors in the spreadsheet used in their seminal work, This Time it’s Different, to demonstrate what is perhaps their most famous point (albeit one I didn’t refer to in my piece), namely that debt/GDP levels above 90% have a significantly negative impact on GDP growth. The article pointing these errors out, and to which Blanchflower alluded, implies that debt can reach higher levels than 90% (and accordingly higher than those currently recorded) without damaging growth. Therefore the fiscal charter is guarding against a danger that doesn’t exist.
- Government investment can generate high returns which justifies running a budget deficit even in good times.
1. Higher debt levels
What’s the fuss?
No one denies that debt rises, whenever there is a recession, for purely cyclical reasons (tax take drops with the economy and spending rises as more people lose work, need benefits etc.). Therefore my argument was that if you can’t stop debt rising in the bad years, and on top of that you choose to allow it to increase in the good years, then debt will just keep on going up, inexorably, without interruption. I think that’s dangerous. But if Blanchflower and Herndon et al. are right, then Reinhardt and Rogoff are wrong to claim that debt levels above 90% cause serious damage to growth. And if debt/GDP can go over 90% without a negative impact on GDP, then maybe there is no reason for debt not to keep on increasing all the time. That would mean that the inexorable increase in debt I cautioned against isn’t necessarily a dangerous thing. If data doesn’t show that all countries with high debt/GDP levels have low GDP growth rates, one might argue for letting debt increase indefinitely. Why worry?
Of course, none of Reinhardt and Rogoff’s critics argue that there is no limit, either discernible or hypothetical, to how high debt/GDP ratios might go before they become unsustainable. In a New York Times Op-ed piece Herndon’s co-authors Robert Pollin and Michael Ash wrote: “History suggests that there is some threshold beyond which piling on public debt definitively yields lower economic growth, but there is no consensus on what that threshold is, and the evidence suggests, in any event, that the United States and Europe are not anywhere close to it” (29/4/2013). So there is a threshold out there – somewhere – but until we find it there’s no point in arbitrarily curtailing debt – that seems to be the implication.
I think there is a risk in ever increasing debt and leverage. That risk is intimately linked to the perceived benefits of the (currently much in vogue) policy of government debt fueled expansion. Those in favor of increasing government borrowing assume that debt funds high return projects which increase GDP and therefore enable the government to borrow more and so on – a kind of virtuous circle. The problem is that the longer that virtuous circle goes on, the more damaging the impact if it ever finally stops, which it typically does – trees don’t grow to the sky. There are many reasons why government investment funded by debt will not necessarily lead to increased growth: loss of competitiveness, raw material shock, war, recession in a trading partner, financial crisis, failure of one of the previously productive investments, just to name a few. If that happens, the country is left with the debt but doesn’t have the GDP to support it. In other words, the longer you embark on this virtuous cycle the more you expose yourself to the risk that your investment strategy doesn’t bear fruit – if for example the productivity benefit of shaving half an hour off the journey time between Birmingham and London doesn’t provide a good return on the £80bn cost (mooted by the IEA) of HS2.
There are two main risks attendant on ending up with too much debt relative to your GDP: negative impact on GDP growth from the debt burden (the risk discussed by Herndon et al.) is the lesser of the two. The greatest risk is of course outright default on government debt. The most important data in Reinhardt and Rogoff, which neither Blanchflower nor Herndon contradict, is their record of the frequency and ubiquity of debt crises in general and of government default in particular: in a typical year since 1800, 5-10% (GDP weighted) of all countries in Reinhardt and Rogoff’s broad data set were in some form of external default, with nearly all of the strongest economies having experienced default at least once (UK included). If 5-10% seems like a low number you’re being very insouciant. Default is a damaging and painful event which countries take years to recover from if they ever recover at all. If someone offered you a bet in which you had a 5% chance of poverty for ten years, would you take it?
Table 8.4 in This Time it’s Different shows the growth in both external and domestic debt levels for countries which experienced external default between 1827 and 2003 (89 episodes). The average increase in the four years preceding the default is 35-40%. It therefore seems that a rapid prior increase in debt is typical for countries that default. In Bayesian terms, conditional on having defaulted, the probability of a country having rapidly increased its debt is high. Reinhardt and Rogoff’s data also shows that government debt increases rapidly in the run-up to banking crises, and that housing price increases are a strongly predictive factor for any debt crisis. In other words, default and financial crisis are typically preceded by increased leverage and asset price bubbles.
That does not mean that rapid increases in debt always result in default. There are many countries which increase their debt far more rapidly than the average of the defaulters in Reinhardt and Rogoff’s sample, and yet do not default. That’s because non-defaulting countries are typically 90-95% of the sample, so in any given year most countries do not default. Whatever you do with your deficit, your “Bayesian prior” chances of default are low, which is consistent with the fact that individual countries, their neighbors, creditors and international institutions will go to great lengths to avoid any default – and that paying your debts is the norm. Increasing your borrowing levels rapidly may result in a “posterior” increase in your chances of default (as discussed in the previous paragraph), but this is outweighed in your total probability of default by the prior probability. However, allowing your debt to rapidly increase fulfills one of the essential conditions of defaulting on it. It makes you eligible to be one of the 5-10% defaulters in any typical year. It’s like driving a car at over 100 mph. It drastically increases your chances of crashing, even though most people who drive at that speed do not crash.
In simple terms, this is gravity at work. The virtuous circle assumed by those in favor of increased debt places the countries which pursue it at greater risk of crashing back to earth, whether in the form of default or just (as discussed by Herndon et al.) lower economic growth. The more a country’s debt grows, in apparent lock-step with a GDP which is increased by the investments funded by that debt, the more its economy is exposed to a sudden fall in GDP and therefore to default. The banking sector in the run-up to the financial crisis is an excellent analogy for this: banks grew their loan books faster than their equity base, the companies they lent to expanded and increased their profits and therefore their ability to service their debts, the default risk those companies represented (as borrowers) on the banks’ balance sheet reduced, and the banks (as lenders) therefore seemed highly solvent – despite their low equity ratios. But when sub-prime CDOs, an obscure part of the market, began to collapse, suddenly the whole financial system had to deleverage. The banks’ liabilities were still there, but the value of their assets and their borrowers’ ability to repay evaporated. It was like a soufflé. The banks were left with too much debt, too many poor assets on the other side of the ledger, and not enough equity supporting the whole edifice.
As Warren Buffet said about derivatives in 2005, you learn who has been swimming naked when the tide goes out. Just so, countries which increase their deficit to invest for growth expose themselves to an increased risk that the expected growth will not be achieved or be unsustainable.
Increasing debt too rapidly carries the low probability but high impact risk of negative growth, default or some kind of bail out, such as experienced by the UK in the 70s. So yes, there is no debt/GDP limit which is set in stone; you can try to increase debt/GDP levels above 90% and get away with it. But the more you pursue that approach the more you expose yourself to the risk of very serious financial calamity. If this seems like an unreasonable warning remember Greece’s default seemed unthinkable until late 2009. The most valuable lesson of This Time it’s Different is that these low probability high impact risks never seem remotely possible beforehand to the governments which actually experience them. They always think “this time it’s different” – for them.
Then again, maybe, if you’re lucky, the excessive leverage you take on to fund grands projets, whose contribution to GDP growth does not materialise, won’t be enough to make your country go bust. But – even in this lucky scenario – that borrowing will still leave your country in the uncomfortable situation of being saddled with a higher debt ratio which then mechanically becomes worse due to the pro-cyclical effects of recessions. And in such an adverse scenario the government in question has less fire-power to stimulate the economy. That’s more or less where the UK and other countries were in 2009. Yet Blanchflower said on BBC Radio 4’s Today Programme “there’s certainly some views now that actually what we should probably be doing is not austerity but large amounts of fiscal investment and stimulus to the economy, to essentially prepare it for the shock we know is coming” (14/10/2015). So there’s a shock coming, according to Blanchlower. And history tells us such a shock will cyclically increase borrowing levels. Incredibly, what Blanchflower advocates is going into the potential shock with higher borrowing, to fund “fiscal investment” (by which he means government investment funded by borrowing) which is so productive and high return that it enables the economy to grow so robustly and sustainably that the shock never happens in the first place. In other words we lever up to invest and grow through the shock!
Far from there being any precedents for such a strategy actually working, the data we have on recessions tends to suggest that these tend to be preceded, not avoided, by strong GDP growth. Countries typically endure recessions because they grew too fast, over-extended, and, as described above, over-borrowed. Figure 10.4 in This Time it’s Different shows that real GDP growth of the average non-emerging country entering a banking crisis is over 2%, already a strong figure, with countries experiencing the “big 5” banking crises recording growth of around 5% on average two and three years before.
What Blanchflower is probably envisaging is the government borrowing to invest in infrastructure – not so much to stimulate growth – but to make the country efficient and competitive, allowing it to enter any slowdown in a leaner and tougher state. It’s like doing extra sit-ups in the weeks leading up to a boxing match so you’re better able to withstand the punches in the stomach “you know are coming.” I can testify that this makes perfect sense in boxing, but as outlined above the data suggests that most recessions are not due to supply inefficiencies, but to demand increasing too rapidly and then unwinding. If activity is down, as it was after the last financial crisis, and fewer people are traveling, having lots of quick train connections is not going to help you. The only way to avoid such a collapse in demand is not to over-hype it in the first place. We might have avoided the sub-prime crisis for example if Greenspan had slowed the US (and global) economy by proactively tightening interest rates and – especially – capital requirements for banks. Increasing leverage to invest in infrastructure, on the contrary, increases growth and therefore the risk of future recession.
This sort of thing reminds me of the Dukes of Hazard, when they are chased by Sheriff Roscoe P. Coltrane and his (Captain Pugwash-style double entendre) side-kicks Enos and Cletus, and escape by accelerating fast enough to fly over a parked truck or some other obstacle (usually fitted with a convenient ramp). Somehow, higher growth through public sector investment going into recession will avoid the recession. You can almost hear the boys shouting “Yeehah” and the characteristic jingle of their car, “General Lee”:
But let’s say a Corbyn-style government was able not to conform to these precedents, defy the odds and, as it were, follow the Dukes over the chasm: does that mean levering up into a recession is a good policy? Again, if it works, it’s fine, but the chances of it not working are not zero. All you need is one duff infrastructure investment and suddenly your growth falls below forecast. Instead of driving over the canyon you end up crashing into its depths. You go into the crisis thinking your investments are going to turn you into Germany, and instead you end up like Greece. As I say, what statistics we have show the probability of successfully investing to grow through a recession is low. But even if it were high, the impact of failure is disproportionate. The benefits of nine years of Germany can be more than wiped out by one year of Greece.
What Blanchflower might say is, “o.k. if it doesn’t work we’ll just borrow even more to get ourselves growing again.” But this simply exposes the economy again to the same risk of the debt funded investments not working, but this time starting from a more indebted position. The same risk, just more of it. It is a double or quits strategy. History shows us that governments are exceptionally reluctant to believe that their limit at the casino will run out, and prone to believing that their next bets – errr, I mean infrastructure investments – are going to work. They are therefore exceptionally prone to exposing their country to the risks of such doubling up.
Fooled by randomness
The 90% cut off point in This Time it’s Different was the part I found least persuasive. The reason for that, quite apart from the errors spotted by Herndon et al., is that GDP growth is affected by a vast array of fundamental factors – demographics, globalisation, technology and raw material prices to name but a few. Not only are there many factors, but each of these factors affects different countries in different ways at different times. The same limitations affect any single factor analysis of default risk. There is also a boring short term cyclical factor when analysing GDP growth: you can have very high growth in one year because your economy was depressed before, and negative growth in another because your GDP previously reached an unsustainable peak; any calculation which averages isolated observations of these cyclical ups and downs can be capturing a lot of statistical noise (as implied by Herndon et al.).
Calculation of both default risk and GDP growth therefore requires multi-factor models. And it is impossible to isolate the impact of any of these factors on GDP growth or probability of default with certainty. You can only observe the GDP growth or the default. To isolate a cause, you would have to re-run history to see what GDP growth was in a particular case with this or that factor altered, or whether a country defaulted in a particular case absent this or that factor. Of course we have the 650+ years of default experience, but the world changes all the time, so a default in 1340 does not take place in the same environment as one in 1815. Looking at a history of defaults, however long, is not the same as a controlled experiment.
Policy decisions obviously can have an impact on our probabilities of default. Yet we can only observe the outcome of those probabilities (i.e. did we default or not) not the impact of our policies on those probabilities (i.e. did our probability of default increase to 14% or decrease to 6%). In other words, we can pursue a bad policy which increases our default chances from 30% to 60%, and still have a 40% chance of getting lucky and avoiding default. Similarly, we can reduce it to 5% with prudent policies and something from left field can still put us in debtors’ prison. But the only thing we can actually observe is the non-default in the first case and the default in the second, not the 60% and 5% respective ex ante probabilities of default.
Nassim Taleb might say that those who, like Herndon et al., look at countries with high debt/GDP levels which experienced high GDP growth, or those like Reinhardt and Rogoff who – from a different perspective – point to countries with high debt/GDP levels which experienced low GDP growth, are fooled by randomness (to quote the title of his brilliant book first published in 2001). They only look at the outcomes they observe together with a few potentially explanatory factors, without taking into account the fact that those factors might have resulted in dramatically different outcomes and therefore dramatically different statistical relationships, cut-off points etc. The analysis of the factors we cited above, such as a rapid increase in debt levels or rapid economic growth, only tells you that those factors increase your likelihood of default. They can’t tell you when you are going to default or why.
So we can conclude that:
(i) defaults and growth shocks are possible
(ii) they are extremely damaging
(iii) any data we have about their occurrence is a poor guide as to their probability of occurring in the future, so we can never be sure when they may occur and
(iv) government investment does not have a 100% chance of sustainably increasing GDP,
(v) probability of default or growth shock seems to increase with increased leverage and
(vi) there are good reasons for this increased probability (the laws of gravity).
In other words we know that by increasing borrowing we increase our risk of financial distress, we just don’t know when that risk might materialise. It’s like playing Russian roulette – without knowing how many bullets are in the gun.
It seems incredible therefore that Blanchflower uses Herndon et al.’s rebuttal of Reinhardt and Rogoff’s 90% cut-off point as an argument for more borrowing. If you knew where your cut-off point was, you could comfortably leverage up within that limit and know you weren’t taking an excessive risk. What Herndon’s paper shows is that we really don’t know where that limit is for any given country at any given time. His co-authors acknowledge, as indeed would anyone but a lunatic, that there is a limit out there – somewhere – they’re just not sure where that limit is. For the UK at the current moment it may be 110%, but then it may also be 85% and we just don’t know it yet! If no one knows where the the critical point is, it’s best to be prudent. If you’re driving on a road with poor lighting and can’t see around the bend you don’t use that as an excuse to floor the accelerator! Unless you’re Bo and Luke Duke of course. Yeeeeeehah!
2. Productive government investment
“Valuable investments that will give a return of more than 2%”
Also on the Today Programme of 14/10/2015, Stephanie Flanders argued against the fiscal charter in these terms:
We have interest rates extremely low, … the cost of the government’s borrowing is lower than 2% at the moment, if a government can’t come up with valuable investments that will give a return for future generations of more than 2% a year, that doesn’t seem to be a very imaginative government.
According to this much feted economics commentator, who once had a show on BBC Radio 4 called “Stephanomics” (illustrating, some might say, that station’s general ineptitude in financial matters), the fact that governments can borrow at less than 2% means that any investments they make which return more than 2% are potentially valuable investments. Valuable investments that return 2% a year actually sounds like an oxymoron to me, and reminds me of a very funny episode of the 90s sitcom Friends where the “friends” fall out over lottery tickets which they bought as a syndicate, and then taunt each other with what they’ll do with any winnings they get from playing on their own. Ross, square as ever, exclaims: “If I win, I’m going to put it all into a very low yield bond!” (starts 2.10 minutes in).
If the government can borrow at 2% and invest at a 3% rate it can make a 1% spread. You don’t need to be an economics correspondent to understand that. The problem is – and it’s a spectacular one coming from someone ostensibly working for an organisation purporting to manage people’s savings – that you can’t make an apples with apples comparison between the rate at which the government borrows and the rate of return at which it invests without taking risk into account. If the government issued treasuries at 1.8% to buy corporate debt yielding 5%, Steph might say “yay, a 3.2% spread, ace!” However corporate debt carries much higher default risk than the government’s, and after adjusting for that risk it isn’t necessarily much of a spread at all. Of course taking risks such as these can be a good thing for people with expertise in such matters, such as (at the, cough, risk of contributing another oxymoron to the debate …) competent professional bond investors.
The question is how good governments are at assessing such investment risk. Clearly they don’t typically invest in corporate debt (although central banks have been doing so recently as part of their quantitative easing measures), rather they buy what fiscal expansion supporters think will be solid, productive assets. Real, worthy assets, not financial investments like corporate bonds. In the twitter discussion in response to my piece, advocates of fiscal expansion came up with case studies of productive government investment which were interesting to say the least. Astonishingly for me, Blanchflower’s first and only specific example was the channel tunnel, which though clearly a productive investment initiated by two governments, is a curious example to choose today because it happened so long ago.
In fact most people on the thread went even further back in time, citing the Norman conquest, Roman roads and to cap it all off – the pyramids!
Supporters of fiscal expansion also came up with a few potential government projects which might enhance GDP in the current environment, all of them public transport investments with an emphasis on railway.
While they are different from corporate bonds, and may yield more than 2%, even such real assets have many project specific risks, over and above the macro-economic risks discussed above in the first section. There is construction risk, with most large government projects going significantly over budget and taking longer to complete than expected, and demand risk, e.g traffic assumptions failing to meet forecasts. Such mistakes are described by Kahneman and Tversky as the “planning fallacy”; Bent Flyvbjerg, a seasoned researcher into mega-project cost over-runs, provides an excellent reading list for students of the planning fallacy.
Although not all governments will be as insouciant of this risk as Flanders, the literature shows that, far from being immune to the planning fallacy, governments are the providers of its most egregious examples. Remember that the latest completed large public infrastructure work in the UK is the Edinburgh trams. Crossrail, still ongoing, was £6.6bn over budget by 2014 [update (27/12/18): Crossrail hit by further budget over-runs and delays; update (3/4/2019): Crossrail delayed to 2020 “at the earliest” and cost increases again, from £14.8bn to £17.6bn]. Advocates of infrastructure investment funded by government borrowing show little awareness of this risk.
Many large infrastructure projects fail because they are overtaken by technology. At the moment, numerous large oil drilling projects are no longer viable because the price of oil has fallen, due to the twin technological innovations of fracking on the supply side and energy efficient cars and planes on the demand side (U. S. oil demand in boe has been flat to down during the last three years of strong economic growth in that country). Although it is too soon to say that we will all be working from home in virtual offices, anyone looking to trains as a cast iron high return investment should pay some attention to innovations like Skype, logmein and Gotomeetings which reduce the need for face to face interaction to do business.
The Japan precedent
It’s not for nothing that twitter supporters of government infrastructure spending went so far back in time. The value of these sorts of grand projets is that they seem to last forever. Once you build one, you don’t need another one for ages, if ever. There is no need to build a second channel tunnel or another M6 or another East Coast mainline. The more you build, by definition the less there is left to build. And, rationally, you tend to start with the most needed projects, and once you’ve finished what’s left is, although perhaps important, less crucial. That’s why the 80s had the channel tunnel, and by contrast 2015 only has HS2 as poster boy for productive government investment. Japan of course acted as though it could ignore this law, and is now regularly cited as an example of how a country can run out of useful public works, and reach a point where incremental infrastructure projects add nothing whatsoever to GDP growth. Instead you get repeated and pointless re-tarmacking of roads no one drives on. Of course the UK is far from that situation, and its train service is far from that of the Japanese bullet trains, but advocates of government borrowing to fund infrastructure investments should remember the inevitably decreasing incremental returns from such investment.
The dangers of out of sample forecasting
The most dangerous part of Flanders’ remark is her use of the government’s current borrowing rate to evaluate infrastructure projects with a lifetime of over 20 years. This tendency is very widely spread. The fact that borrowing rates are so low somehow makes many people think that infrastructure investment is a no-brainer – “it just needs to return more than 2%!” Not many things can be guaranteed in life, but I can guarantee that sovereign interest rates will not stay below 2% forever. Even if a project makes its target return, any government investing in that project is stuck with those returns for the life of the project. In the meantime it has to pay interest on its debt at whatever rate the lenders are willing to accept. So the return on the asset is capped, but the payment on the liability is not. Even if the government issues ultra long term debt with the same maturity as the asset, at some point the debt has to be refinanced. If the asset’s impact on GDP is a 2.5% increase, and interest rates have gone to 5% by the time the debt is refinanced, investment in that asset has caused a sharp deterioration in the government’s solvency – despite the fact that it returned more than the government’s cost of borrowing at the time of the initial investment.
Nate Silver in his brilliant The Signal and the Noise (2012) attributes Moody’s failure to anticipate the credit risks exposed by the sub-prime crisis to the impact of “out of sample” events on forecasting. Essentially, the out of sample mistake involves using data from a particular period to forecast events whose properties make the period irrelevant to them. Thus (in keeping with our Dukes of Hazard analogy) someone might justify driving home drunk by the fact that they’ve driven for years without a single accident. The problem is that the driver was sober in that period, so the data from his years of sober driving are not relevant to his safety while driving drunk. Silver says: “Moody’s estimated the extent to which mortgage defaults were correlated with one another by building a model from data … going back to about the 1980s. The problem is that from the 1980s through the mid-2000s, home prices were always steady or increasing in the United States.” The use of the current depressed interest rate to evaluate long term investments such as infrastructure would expose the governments making those investments to classic “out of sample” risk.
This may seem like a theoretical consideration from an economics textbook … but only if you haven’t been paying attention to the last financial crises we have experienced. All of them started when a particular region or sector benefited from an abnormally low cost of capital, and all of them bloomed when that low cost of capital was extrapolated into the future – as Flanders does with current government borrowing costs – and a bubble ensued. The Asian crisis of 1997 had its roots in ultra-low rates of dollar borrowing in those countries premised on a perceived reduction in country risk and the perpetual strength of the “Asian Tiger” economies, leading to massive over-expansion and over-leverage which collapsed disastrously when investors lost confidence in Asian currencies. The same is true, minus the Tiger metaphor, for Russia in 1998. The dotcom bubble at the end of the 90s was facilitated by astronomic price/earnings ratios for tech companies which enabled them to raise vast sums cheaply through equity issuance, and led to massive over-investment in semi-conductor capacity and fiber optic cable among others.
The sub-prime crisis’ inception came at a time when Asian countries’ (mainly China’s) trade surpluses with the US were recycled into US treasuries to prevent their currencies from appreciating against the dollar, depressing US government borrowing costs relative to what they should have been given the strong economic activity underpinning such voracious importing of Asian goods. The impact of low government bond yields was amplified by low equity capital ratios (enabled by the financial regulatory establishment in general and Greenspan in particular) at banks and insurance companies (which guaranteed much of the debt issued in the period) through the multiplier effect. This easy money was extrapolated into the future, leading investors to buy ever more risky and opaque investments like sub-prime CDOs, resulting in a massively levered exposure to highly toxic assets (I write about this in detail here). And the European sovereign debt crisis had its genesis in the convergence of the borrowing rates for high risk countries like Greece and Spain with those of lower risk countries like Germany. Again, this was extrapolated into the future, and guess what, a massive infrastructure boom ensued, particularly in Spain, where much of the infrastructure is currently under-utilised.
Using low funding costs as a reason to invest is the siren song which unites all of our recent financial crises. Flanders’ casual obliviousness to this fact is a perfect illustration of how human beings are primed to forget the lessons of past financial crises, and why the fiscal charter is such a good idea. To paraphrase Flanders:
If an economist can’t come up with plausible scenarios in which future generations will be confronted with borrowing costs of more than 2% a year, that doesn’t seem to be a very imaginative economist.
You don’t need the government to invest
Another problem with the fiscal expansionist thesis is that it often neglects the fact that investment does not need to be carried out by the government. Of course, very few people actually think that only the government can invest, but both Blanchflower and Flanders do argue that if the government cannot borrow this will limit the ability of the economy as a whole to invest. This is a very blinkered perception. We are currently benefiting from a deluge of products which are the result of heavy investment which has nothing to do with governments: think smart phones, social media like instagram and pinterest, intermediation channels like Uber and AirBnB, free telephony like WhatsApp and Viber, electric cars from Tesla, discount supermarkets like Aldi and Lidl, discount airlines like Southwest or Ryanair.
In infrastructure, the London Gateway container ship port was privately built by DP World and will eventually have capacity of 3.5m TEU (equivalent to the capacity of 20 of the largest container ships in the world). Its impact on the efficiency of transporting goods into the UK is massive. The most pressing infrastructure requirement in the UK is currently airport expansion, and that will also be privately funded, wherever it happens (if ever it happens). The roll-out of fibre in the UK is largely funded by BT Openreach. This may not be to everyone’s liking, but it is a bit rich to say that we need the government for productive investment when there is so much obviously productive private investment happening right under your nose.
The government doesn’t need debt to invest
Having said all that, there are definitely some productive investments which only the government could make, only a fool would deny that. The question is, does the government really need to increase its borrowing to fund them? Take HS2, one of the most expensive future train journeys know to man. At an upper level cost of £80bn over an ambitious 10 years it will cost £8bn per annum. Much more likely it will take closer to 20 years, implying £4bn per annum. £4-8bn p.a. on the largest and most expensive infrastructure project both the Labour and Conservative party can come up with is equivalent to 0.5-1% of annual government expenditure. In other words, 1% GDP growth pays for another HS2.
In summary, the fiscal expansionist strategy is one which exposes the economy to the risk of bankruptcy, all for the benefit of … valuable investments yielding over 2%.