Inflation not the easy way to cut public debt burdens - Capital Economics
Global Economics

Inflation not the easy way to cut public debt burdens

Global Economics Focus
Written by Vicky Redwood

One of the key consequences of the coronavirus is set to be a sharp rise in government debt burdens. We have already looked at whether governments can just tolerate their rise in debt here. But this is not their only option. Default is one, austerity is another, but in this Focus, we look at the third – namely inflating the debt away. We start by asking why and how governments would go about this. We then look at whether it would work in practice, before finishing with where we might see it attempted.

  • Inflation is the not the easy fix for the coronavirus-related rise in public sector debt that it might seem. Not only might it not actually do much to reduce public sector debt ratios, but higher inflation would impose other serious costs on economies. Governments are best off reducing debt burdens through other means if they can. Where higher inflation does happen, it is likely to be either accidental or pursued out of desperation.
  • There are three key reasons why high inflation would not be as helpful in reducing debt burdens as is often assumed. First, a rise in inflation would push up new borrowing costs, making it more expensive to finance deficits and refinance maturing debt. There would still be an initial drop in the debt ratio, given that most government debt does not mature straightaway. However, the average maturity of government debt is not that long, plus we expect most countries to be running a budget deficit over the next couple of years.
  • The bond market reaction would therefore be crucial. In a favourable scenario, markets might take a while to react (as historically they have done) and would also believe that the rise in inflation was only temporary. But given the difficulties of raising inflation by a fixed amount for a set length of time, there is a risk that markets would instead doubt that inflation would be brought under control again. Indeed, if bond yields rose by more than inflation, inflation could even up raising a government’s debt burden eventually.
  • Governments could seek to sidestep these issues with the bond market, either by monetising the debt and/or by financial repression (i.e. forcing the private sector to buy debt at below market prices). But these strategies would also have severe drawbacks and could risk letting inflation get completely out of control.
  • The second key problem is that high inflation may harm real economic growth. Not only would this be undesirable in and of itself, but it would also lessen the drop in the debt to GDP ratio. The process of raising inflation could also cause problems. Indeed, one of the key problems in all this is that raising inflation is hard and imprecise. With aggressive enough policies, it could be done. But trying to raise inflation through large amounts of QE, for example, would risk inflating another asset price bubble which, when it burst, could prompt another financial and economic crisis.
  • The third problem relates to bringing inflation back down again, which would probably require a sharp slowdown in the real economy. Of course, governments could simply choose to live with high inflation once the debt burden had been reduced, but this would also inflict significant long-term damage on the economy, by reducing investment and distorting price signals.
  • For all these reasons, when inflation has been used to reduce debt in the past, it has usually happened because a government has resorted to this out of desperation and weakness, rather than because it has judged that it is in the best interests of the economy in the long term.
  • Fortunately, most countries can just tolerate their high levels of debt – meaning that they can, and will, avoid going down the inflation route. Even countries where debt is not on a sustainable path might not try to inflate away their debt either. For some, including the individual euro-zone countries, higher inflation might not be an option, while for others, including Brazil and Mexico, the alternatives of austerity combined with financial repression might be preferred. But if this does not work, then some of these countries may resort to inflation further ahead. So, too, might Italy if it were ever to leave the euro-zone.

Inflation not the easy way to cut public debt burdens

One of the key consequences of the coronavirus is set to be a sharp rise in government debt burdens. We have already looked at whether governments can just tolerate their rise in debt here. But this is not their only option. Default is one, austerity is another, but in this Focus, we look at the third – namely inflating the debt away. We start by asking why and how governments would go about this. We then look at whether it would work in practice, before finishing with where we might see it attempted.

The incentive to inflate

Government debt as a share of GDP over the next year or two is already set to rise by between 5 and 15 percentage points (pps) in EMs and between 15 and 30 pps in DMs. The rises could be bigger if the virus returns in widespread waves or if economies are slower to bounce back than we expect.

Whenever government debt has risen sharply in the past, including after the global financial crisis, the question of “just” inflating it away has reared its head. This is not surprising; on the face of it, it seems like an easy way out for governments. Much of their debt is fixed in nominal terms. So other things equal, the faster the price level and therefore nominal GDP rise, the bigger the drop in debt to GDP ratios.

It is also possible that, by boosting tax revenues, higher inflation would help to reduce the deficit, therefore also reducing the speed at which debt rises. That said, given that inflation would also raise many sorts of government spending (including benefits automatically linked to the rate of inflation), this would not be guaranteed. So in this Focus, we assume that the net effect of inflation on the primary deficit is neutral, meaning that if inflation is to have any impact on the overall deficit, it has to be through its effect on debt interest.

Raising inflation would have advantages other than reducing the real value of public sector debt. For example, it would reduce the real value of private sector debt too (although we come back to this point later). And higher inflation, and thus higher nominal interest rates, would give central banks room to cut rates further during the next downturn. Here, though, we concentrate on whether inflation can help governments to address their soaring debt ratios.

But why would governments even consider going down this route when, as we explained in our previous Focus, most of them can tolerate higher debt burdens? Well, we also explained that not all countries are in this position. And even those that can sustain their higher debt may nonetheless be tempted by the quick fix of higher inflation. After all, high levels of debt in some of these countries will only continue to look sustainable if interest rates stay low. Moreover, waiting for economic growth to reduce the debt burden could take decades; higher inflation, on the other hand, seems to offer a much quicker reduction, as we shall see below.

Meanwhile, for countries where a high share of their public debt is owned by foreigners (and as long as this debt is denominated in the country’s own currency), domestic creditors would be spared some of the burden of higher inflation. For example, 35% of marketable US government debt is owned by foreigners. As inflation rose, we would expect the nominal exchange rate to fall, reducing the foreign currency value of the debt and its repayments. (Although if all countries tried to inflate away their debt, this effect would obviously net out at a global level.)

Let us start with a deliberately simplified stylised example to show how inflating away the debt could, in theory, work. We assume an initial virus-related jump in the public sector debt ratio from 100% to 120%. Our base case assumes that debt as a share of GDP then stabilises at 120%. This is based on a balanced primary budget and nominal interest rates equal to nominal GDP growth of 4%, which in turn reflects real growth of 2% and inflation of 2%.

Chart 1: Stylised Example of Government Debt as a % of GDP (Interest Costs Held Constant)

Source: Capital Economics

Now imagine that inflation rises, while real GDP growth and interest costs are the same as in the base case. Interest costs might stay the same because the government has perpetual debt that pays a fixed interest rate forever, or because market bond yields do not change when inflation rises. Inflation of 4%, rather than 2%, would take debt as a share of GDP back to 100% within 10 years. (As an aside, that is the same profile as would be seen in a country with nominal interest rates 2% below nominal GDP growth – which, as we showed in our recent Focus, includes many of the major economies.) An inflation rate of 6% would do it in five years; an inflation rate of 10% in less than three years.

But how?

One issue to address early on is how governments would actually engineer such a rise in inflation when most countries these days have independent central banks with clear inflation-targeting regimes. This is not just true of developed markets (DMs); most emerging markets (EMs) now also have independent inflation-targeting central banks. One option would be for governments to revoke central banks’ independence in order to set inflation rates themselves for their own gain. But that seems unlikely, certainly in all DMs and even most EMs too.

Nonetheless, a rise in inflation could still be achieved by tweaking central banks’ existing inflation-targeting frameworks. After all, the current frameworks are not set in stone; there have been frequent changes in the past, including what measure of inflation is targeted, the exact level of the target etc. There are various possibilities which could be justified as improving the inflation-targeting regime but would result in a helpful rise in inflation for the government.

One is simply to raise the inflation target. There has been plenty of discussion in recent years about doing this to get around the problem caused by the effective lower bound on nominal rates. Another option would be for the inflation target to be extended to a fairly wide range. We have argued before that this might sense in any case given that the apparent breakdown of the past relationship between activity and inflation has called into question the ability of central banks to control inflation. (See here.)

Alternatively, central banks could adopt a price-level target – again, something which has also been widely discussed in recent years. This would mean central banks pledging to make up for any policy misses in the following years. It, too, could be justified by helping to avoid the problems caused by the zero nominal bound; in theory, if inflation falls short of the target, inflation expectations rise, making the current policy setting more powerful in real terms. And given that inflation has in recent years undershot the target more than it has overshot it, this would give central banks the scope to run inflation above its current target for a period. That said, these undershoots are generally not big at the moment; in the US, for example, the level of consumer prices is 2.5% below where it would have been had inflation been at the target since 2009. (See Chart 2.)

Chart 2: US CPI (Jan. 2009 = 100)

Sources: Refinitiv, Capital Economics

Finally, central banks could just turn a blind eye to any inflation overshoots that occurred within the existing inflation-targeting framework. An inflation overshoot could happen if, for example, the inflationary effects of the policy stimulus provided during the coronavirus crisis were more powerful than anticipated. Indeed, we shall discuss later whether an unexpected rise in inflation like this could be more beneficial for the debt ratio than a rise that was flagged in advance by a change in the central bank’s target or mandate.

Note that there have been many previous episodes where inflation has persistently either overshot or undershot its target for several years. For example, inflation exceeded the target in the UK for the whole of the four-year period from 2009 to 2013. Meanwhile, Japan’s inflation rate has been consistently below its 2% inflation target since it was introduced in 2013.

Admittedly, any changes to the inflation target would not necessarily be problem-free. (See here for some discussion of the issues.) Governments would need to weigh up the disadvantages of a target change versus the advantage of facilitating higher inflation. Nonetheless, the point is that the independence of central banks and the current set-up of inflation-targeting need not be an insurmountable obstacle to a government trying to inflate away its debt.

Borrowing costs would rise

But that does not mean that attempts to inflate away the debt would work. We set out a simplistic scenario in Chart 1; now we will discuss how the favourable assumptions that we made are unlikely to hold in practice.

One of the main problems is that interest costs would not hold constant as inflation rose, for three reasons. The first is that, in many countries, some of the government debt is directly linked to inflation – meaning that both the principal and the interest payments would rise in line with inflation. The share of index-linked debt is actually quite low in most major developed countries. (See Chart 3, which shows the main countries which issue index-linked debt.) But the UK is an exception, with 27% of its outstanding debt index-linked. And some of the major EMs have similarly high shares.

Chart 3: % of Government Debt that is Inflation-Linked

Sources: National Treasury Offices

The second complication, for countries that have debt issued in foreign currency, is the impact of higher inflation on the exchange rate. Raising inflation would generally – unless accompanied by a sharp rise in official interest rates – lead to a drop in the exchange rate. The value of any foreign currency debt and accompanying interest payments is fixed in foreign currency. But as a result of the lower exchange rate, the domestic currency value of the debt and interest payments would rise along with nominal GDP, leaving the debt burden unchanged. In fact, if the exchange rate depreciated ahead of any rise in nominal GDP, there could be a period in which the ratio of debt to GDP rose. That said, few countries issue debt in foreign currency these days; all DMs and even most EMs issue their sovereign debt in domestic currency.

The third and biggest problem affecting all countries is that any rise in inflation will tend to push up the cost of new borrowing, as investors demand a higher interest rate to compensate for the higher inflation rate. Higher inflation thus makes it more expensive to finance any deficits the government is running, given that deficits will obviously require borrowing new funds from the market at the prevailing rate. And it will also increase the cost of rolling over any existing debt that matures and needs refinancing.

In Chart 1 we assumed that interest costs remained constant; now let us explore for a moment the opposite extreme in which interest costs adjusted immediately. This would happen if interest rates rose as soon as inflation rises and the government was doing all of its borrowing in the overnight markets, so that the interest its paid on its debt was whatever the market rate was on that day. Then raising inflation would have no benefit whatsoever. In fact, if bond yields rose by more than inflation, the government would be worse off.

This is an extreme example. In reality, bond markets might not react like that and, even if they did, government debt is of longer than overnight maturity. So higher inflation would lead to an immediate, once and for all, reduction in the debt burden as a share of GDP.

But assuming that market rates did in time follow inflation up, inflation would not continue to erode the debt burden indefinitely. As more of the debt matured and had to be refinanced at higher bond yields, the debt of GDP ratio would eventually revert to rising at the same rate as in the base case (although the level of debt would be permanently lower).

The exact path of the debt ratio would depend on three factors:

  • the response of market interest rates;
  • the maturity schedule of the government’s debt;
  • and the amount of new borrowing (the deficit).

If countries had a very long maturity of outstanding debt, and little new borrowing, then it would not matter much if market interest rates rose. Unfortunately, in most countries, that is not the case. The average maturity of outstanding debt is generally around only six or seven years. (See Chart 4.) The UK is the exception, with an unusually long maturity of 15 years. That would put it in a better position to inflate away its debt than other countries, were it not for its high proportion of index-linked debt that we pointed out earlier.

Chart 4: Average Maturity of Outstanding Government Debt (Years)

Source: IMF

Admittedly, governments might skew their issuance of new debt towards long maturities to lock into low borrowing costs. Note that the US Fed has recently announced plans to shift its issuance from short-term bills to long-dated bills. (See here). Indeed, governments may even end up issuing perpetual pandemic bonds, similar to the perpetual bonds issued by the UK to finance the Napoleonic and First World Wars. But any such move will feed through only gradually to the average maturity of total debt.

Meanwhile, we expect all major countries still to be running large budget deficits in 2022. (See Chart 5.) This is partly a result of the coronavirus. We think that economic recoveries will be slow, meaning that automatic stabilisers will still be operating. And there will be pressure on governments to increase spending on health permanently, as well possibly to expand the role of the state more generally.

Chart 5: Government Budget Balance in 2022 (% of GDP, CE Forecast)

Source: Capital Economics

Note that Germany is in a better position than most to benefit from inflating away its debt, if it chose to go down that route. Although the average maturity of its debt is about average, its budget deficit is relatively small and it has a small amount of index-linked debt. However, its aversion to inflation means that it is also the country least likely to go down this route in the first place.

A closer look at the bond market reaction

With debt maturities relatively short and countries set to keep running deficits, the response of the bond markets would therefore be crucial, so let us look at that in a bit more detail.

This response would depend in part on how the rise in inflation was done. In particular, it would matter whether it was flagged in advance (for example, via a formal change in the central bank’s inflation target) or was a surprise (for example, if inflation rose unexpectedly and the central bank just tolerated this overshoot).

A surprise rise would result in inflation increasing before bond yields, meaning that inflation would have a more beneficial impact on the debt burden in the short run. But the added uncertainty created when firms and households realised that the government was “secretly” raising inflation might then result in a sharper than otherwise rise in bond yields – preventing any benefits from higher inflation, other than in the very short term.

In contrast, if a government were to announce in advance its intention to raise inflation, bond yields would probably rise before inflation did. This would reduce the potential benefits from inflation in the short run. But if the government reassured bond markets that inflation was unlikely to keep rising, this strategy might limit the extent of any rise in yields.

Indeed, it would matter whether policymakers announced that any rise in inflation would be temporary or indefinite. A commitment to raise inflation by a limited amount for a finite period would stand the best chance of containing any rise in inflation expectations and bond yields, particularly at long maturities. Not only would it reassure markets that inflation would not keep rising to ever higher rates, but it would contain any rise in the risk premium demanded for the increased uncertainty about the inflation outlook.

However, the markets would need to believe what the government said. Moreover, a limited and finite rise in inflation is not very likely anyway, given that it would not put much of a dent in the debt burden. Going back to our stylised example, imagine that governments raised the inflation target from 2% to 5% for two years and then brought it back down to 2% again. Even if bond yields did nothing, this would only cut the debt ratio from 120% to 114%. (See Chart 6.) Taking a decent chunk off the debt ratio would involve inflation rising either much higher, or for longer. For example, keeping inflation at 5% for five years would result in a bigger (although still not spectacular) 15bps drop. Yet the bigger or longer the rise in inflation, the harder it is to imagine bond markets staying calm.

Chart 6: Stylised Example of Government Debt as a % of GDP (Interest Costs Held Constant)

Source: Capital Economics

It would be hard to achieve a controlled rise in inflation for a limited time anyway. First, there is the problem of hitting the new rate of inflation in the first place. We have seen in recent years the trouble that central banks have had hitting an exact rate. The sheer unpredictability and uncontrollability of currencies also undermines the idea that the authorities could opt for a small, contained rise in inflation. What if the central bank aimed for 5% to 6% but ended up with inflation for 7% to 8%? We imagine that the markets might quickly start to lose faith in the government’s promise to raise inflation by a limited amount for a limited time. In addition, as we discuss later, bond markets would be aware from past experience that getting inflation back down again isn’t easy.

Two things can be learnt by looking at the bond market response to the inflationary episodes in the 1970s and 1980s. The first is reassuring for any government thinking about inflating away its debt. It is that bond markets were generally slow to cotton on that the rise in inflation marked a transition from an era of low to high inflation, and that they were too influenced by the low inflation rates that had prevailed in the preceding decade. Indeed, real yields turned negative for periods of time. (See the period before the dotted line in Chart 7.) We might see the same again, given that financial markets have short memories and it is a long time since high rates of inflation were seen.

Chart 7: US Bond Yields & Inflation (%)


The second is more concerning, namely that when bond yields did finally rise, they stayed high for a long time, even after inflation itself was brought down again. (See the period after the dotted line.) Again, bond markets were too influenced by recent experience and investors refused in believe in the new commitment by central banks and governments. Were markets to over-estimate future inflation now, government could find themselves ultimately gaining little or nothing from raising inflation.

So if bond yields did rise, how would this affect the trajectory for the debt ratio? In our stylised example in Chart 1, we showed that raising inflation from 2% to 4% would – as long as bond yields were unchanged – reduce the debt ratio by 20pps after a decade and 40pps after two decades. Now let us assume that bond yields instead rise in tandem with inflation, and that the government’s debt has the same maturity profile as that of the US now (shown in Chart 8.)

Chart 8: Breakdown of US Government Debt by Maturity (As a % of Total)

Source: US Treasury

The light grey line in Chart 9 shows that there is still some reduction in the debt ratio, reflecting the fact that some of the debt is fixed in the near term. But as more debt matures, the benefit from higher inflation fades. The end result is that the debt ratio after a decade falls by 7pps, a third as much as if yields stayed the same.

Chart 9: Stylised Example of Government Debt as a % of GDP Incorporating Rise in Bond Yields

Source: Capital Economics

Let us continue to assume that i) the debt has the same maturity as that of the US government and ii) bond yields follow inflation up straight away. Then inflation would need to increase to 8% to achieve the same 20pp drop in our initial scenario involving a rise in inflation to 4% but no rise in bond yields. And if the government wanted to limit the timeframe of the rise in inflation to, say, five years, the rise would need to be 10%. Two years, and it would need to be about 20%. Inflation has got to double digit rates in the US and other countries before, but, as we shall discuss later, only with significant costs. Needless to say, the shorter the maturity of a government’s debt, the higher that inflation would have to be raised to reduce its debt burden.

Finally, imagine a scenario where bond yields rose by more than inflation. We assume that inflation rises from 2% to 4%, but that bond yields rise from 4% to 7%. There is still an initial drop in the debt ratio, again reflecting the fact that the majority of debt does not mature in the near term. However, once all of the debt has rolled over, the debt ratio is higher than if the government had never raised inflation at all. (See Chart 10.) The only way that a government could get around this would be to try staying one step ahead of bond investors, by generating higher and higher inflation rates. Result: hyperinflation.

Chart 10: Stylised Example of Government Debt as a % of GDP Where Bond Yields Rise Above Inflation

Source: Capital Economics

Getting around this

Admittedly, there are potentially two ways to get around this problem with the bond market reaction. The first is to monetise the debt, i.e. for the government just to sell its debt to the central bank. In that case, high interest rates on bonds would cause no problems at all. The government would simply pay out higher interest payments on its bonds with one hand and take the higher interest income back from the central bank with the other. At the extreme, the central bank could supply all of the government’s financing and refinancing needs. This would simply represent an extension of the current policies of QE.

But avoiding the problem of high debt interest in this way itself comes with costs. Importantly, it risks losing all control of the inflation rate and prompting a take-off into hyper-inflation. Gone is the option of a controlled minor rise in inflation to 5% or 6% following by a return to normality. This is a world of all or nothing.

The second option is financial repression, namely policies to artificially lower interest rates to below market rates. This could be deliberate, or a fortunate side-effect of other policies such as QE or prudential regulation to improve financial stability. The latter would see pension funds, insurance companies and banks absorb large quantities of government debt below normal market rates. By keeping interest rates and bond yields below the rate of inflation, the government could impose negative real rates on the private sector and therefore enable inflation to reduce the real value of debt interest and the debt burden as a whole.

This avoids the inflationary consequences that might come with monetising the debt. But it is not painless either. It is bound to lead to a distorted and stunted financial system. Moreover, it effectively imposes a tax on all savers and wealth holders.

Damage to real GDP growth

So far, then, we have outlined why the potential impact of higher inflation on the government’s borrowing costs is a serious complication. But it is not the only one.

The second problem is the impact of higher inflation on the economy’s real economic performance. Not only is weaker real GDP growth clearly bad for real incomes and living standards, but it can also undermine the helpful effect that higher inflation has on the government’s debt burden. In other words, the boost to nominal GDP growth from higher inflation could be partly offset by a slowdown in real GDP growth.

There are various ways in which high inflation can dent real economic growth. First, it can reduce the economy’s sustainable rate of growth by interfering with the effective working of the price mechanism (making it hard to discern relative price changes from general ones) and hence impairing the allocation of resources. Second, the higher nominal interest rates and additional uncertainty which usually accompanies them (even if real interest rates are unaffected) tends to diminish investment.

And third, higher inflation can have adverse effects via the private sector debt burden. Of course, high inflation would erode the real value of debt in the private sector in the same way as the public sector. But the private sector would also be adversely affected by the effect of any rise in market interest rates in response to higher inflation. This would push up debt-servicing costs in the early years of the loan – so-called “front loading” – causing a sharp squeeze in disposable income for debtors and potentially provoke a big fall in consumer spending and GDP. Moreover, this would raise the probability of default – threatening financial instability and also making it less likely that borrowers enjoyed the benefits of higher inflation later on in the life of the loan.

As with the public sector, the impact would depend on how much debt was fixed in the short term. Countries vary a lot on this score. Fixed-rate mortgages are the norm in the US, France and Germany. And a rise in their popularity in the UK has taken their share there to about 75%. But in other countries, including China, Australia, Greece, Norway, Portugal and Sweden, variable rate mortgage debt dominates. And even if most mortgage debt is fixed, it might only be fixed for a couple of years, plus at least some credit card and corporate debt will be at variable rates. Moreover, with new companies continuously being formed, and constant turnover in the housing market, there will always be firms and consumers needing to borrow new money – and therefore suffering from any rise in interest rates.

Of course, savers would benefit in the short-term from a rise in inflation (although they would suffer in the long term as inflation eroded the real value of their savings). But because savers tend to spend a smaller proportion of their income than borrowers, the short-term demand effects of higher inflation would still be likely to be negative.

It is worth considering briefly how things would stack up if private sector interest rates did not rise with inflation but remained at previous levels, anchored by near-zero official interest rates. It is possible that this depression of real rates of interest would boost aggregate demand. But it is also possible that people would react to the fall in the real value of their accumulated savings by saving more out of their current income. Indeed, that is what seemed to happen in the UK in 1970s.

For all these reasons, then, high inflation would have some adverse effects on real GDP growth. Admittedly, not all of the effects of high inflation on growth are negative. For example, inflation helps relative prices to adjust without any prices having to fall in absolute terms. Nonetheless, it seems likely that at anything other than fairly low rates of inflation, the negative effects predominate.

The process of raising inflation could also cause problems. After all, one of the key problems in all this is that raising inflation is hard and imprecise. It has been hard enough for central banks to target inflation using the fairly well understood tool of interest rates. And now central banks would have to be using far less conventional tools, in many cases against a backdrop of entrenched low inflation expectations.

To be clear, raising inflation could still be done. Countries that have tried but failed to raise inflation in recent years have probably just not had the appetite for the very bold measures this might require. Indeed, the likelihood that in countries with floating exchange rates the currency would fall sharply on the announcement of a policy of higher inflation fatally undermines the idea that inflation could not be generated. Even in current conditions, there has to be some amount of money that could be pumped into the economy to generate whatever inflation rate you wanted. In the extreme, governments could engage in classic helicopter drops. The problem is that you don’t know what that amount of money is.

Accordingly, there would be huge dangers in deliberately generating inflation. One risk is that inflation would rise much further than intended, resulting in a bigger adverse effect on real GDP growth. Moreover, trying to raise inflation through QE would risk inflating another asset price bubble which, when it burst, could prompt another financial and economic crisis.

Admittedly, the alternatives to trying to inflate away the debt – such as austerity – might also harm real economic growth. The impact of austerity would depend on the types of tax rises and spending cuts. Cuts in public sector investment, or tax rises that blunted incentives, would be more damaging to an economy’s supply potential than cuts in day-to-day spending. Even doing nothing might dent real GDP growth if long-term interest rates rose on the back of concerns about the debt’s sustainability. Accordingly, the damage to real GDP growth that inflation would cause would need to be weighed up against the potential damage from other options.

Getting inflation down again

  • So we have already looked at two problems with attempting to inflate away debt, namely a rise in new borrowing costs and the damage to real economic growth. The third and final problem relates to bringing inflation back down again.
  • As we have already explained, the idea that inflation can be raised by a controlled amount for a fixed period then easily brought back down again is naïve. Lowering inflation again would probably require a sharp slowdown in the real economy. Indeed, the experiences of the 1980s and 1990s recessions show how difficult it has been to reduce inflation in the past, requiring large falls in output and employment. Real GDP fell sharply after the big rises in interest rates seen in both the US and UK in both the late 1970s/early 1980s and late 1980s/early 1990s. On both those occasions, and after the normal time lags, unemployment picked up sharply too. (Chart 11 illustrates this using the US.)

Chart 11: US Inflation & Unemployment Rate (%)

Source: Refinitiv

A further complication is that bond yields would probably take longer to fall than inflation; we showed earlier that this is what happened when inflation was brought down in the 1980s. So there would be a period when, with bond yields above inflation, the reduction in the debt burden would slow or even go temporarily into reverse.

  • Of course, governments could simply choose to live with high inflation even once the debt burden had been reduced – but this would also inflict significant long-term damage on the economy, by reducing investment and distorting price signals.

The past

All of this begs the question of why countries have used inflation to cut their debt burdens in the past.

Part of the answer is that, actually, they haven’t. As we pointed out in our last Global Economics Focus, the major debt reductions by the UK and US in the 19th and 20th centuries were not achieved primarily by high inflation. In fact, in the case of the UK’s debt reduction between 1820 and 1900 after the Napoleonic Wars, inflation did not play a role at all – it averaged around zero. Inflation did play some part in reducing debt burdens after the Second World War. But this was primarily in the years immediately after the war and was largely accidental, in that the high rates of inflation reflected the lifting of the wartime price controls. The reduction in debt seen over the 1950s and 1960s was instead mainly due to the combination of decent real GDP growth and fiscal restraint. Inflation between 1955 and 1970 averaged only 3.6% in the UK and 2.3% in the US. (Chart 12 shows the UK’s experience.)

Chart 12: UK Government Debt as a % of GDP & CPI Inflation

Sources: Bank of England, OBR

Of course, there have been occasions when inflation has worked to reduce the debt burden. This appears to have happened in Brazil and Argentina in the early 1990s, as well as Weimar Germany in the 1920s. German government debt was effectively wiped out by the mid-1920s. However, such episodes have generally happened because governments have slipped into doing it out of weakness, rather than weighing up all the options and judging it to be the rational course of action in the long term. Moreover, as we know from the German experience – where hyper-inflation contributed to the conditions under which the Nazis rose to power – the consequences can be horrible. As Keynes wrote in his 1923 Tract on Monetary Reform, inflation is a tax “which even the weakest government can enforce when it can enforce nothing else”.

Moreover, if anything, it would be harder now for governments to reduce their debt burdens through inflation than it was in the past. This reflects various factors. First, governments now have more debt automatically linked to inflation; for example, the UK only began issuing index-linked bonds in 1981, while the US only started to issue Treasury Inflation-Protected Securities in 1997. Second, the maturity of government debt is shorter now. For example, the average maturity of US government debt after the Second World War was about 10 years (given that long-term securities financed a large proportion of the costs of military operations during the Second World War), compared to six now. (See Chart 13.)

Chart 13: Average Maturity of US Government Debt (Years)

Source: US Treasury

Third, the level of private sector debt is higher, meaning that the front-loading problem we discussed earlier is bigger. And fourth, central banks are generally now independent from governments, so there is more of a hurdle (albeit a surmountable one) to deliberately generating inflation.

Implications for countries

Overall, government indebtedness can be reduced in real terms through inflation, but there are offsetting economic costs which may well overwhelm the advantages from a reduced debt burden.

Accordingly, deliberately unleashing inflation is not the easy get-out that it first appears to be. In fact, it is what countries do when there is no feasible alternative. And fortunately, most counties are a long way from this point. That could change in the future if their economies remained very depressed for an extended period, or if another crisis pushed debt up significantly further. But for now, the low level of interest rates means that most can afford to tolerate their high debt levels and simply let economic growth combined with fiscal prudence erode their debt burdens gradually over time. Even if interest rates were to rise, they would probably deal with it using financial repression rather than inflation.

This fiscal prudence need not even be that restrictive. Let us return to our very first example where the debt ratio initially jumps to 120%, and nominal GDP growth is 4%, which in turn reflects real growth of 2% and inflation of 2%. In many developed countries, interest rates on government debt has been about 2pps lower than GDP growth over the past decade. Were that to continue, then countries could run primary deficits of 3% while still ensuring that the debt ratio stabilised. Or they could run primary deficits of 2% and still get the debt ratio back to 100% within a decade.

Even countries where debt is not on a sustainable path might not try to inflate away their debt either. For some, including the individual euro-zone countries, higher inflation might not be an option, while for others, the alternatives of austerity or even default might be preferred. For example, we think that Sri Lanka will be tempted to default, while the initial response to the poor debt outlook in EMs such as South Africa, Mexico and Brazil is likely to be financial repression combined with austerity. But if this does not work, then it is not out of the question that some of these countries resort to inflation further ahead. So, too, might Italy if it were ever to leave the euro-zone.

Meanwhile, there are a few EMs which have a recent history of debt monetisation and may resort to this pretty soon to finance large budget deficits, which, if done on a large enough scale, may ultimately become tantamount to inflating away their debt. This includes Argentina, Nigeria and Pakistan.

Vicky Redwood, Senior Economic Adviser, +44 20 7808 4989,