Welcome to our latest monthly letter, where we review the macro and market action shaping the investment landscape. This month we are going to discuss the topic of inflation. There's a lot to get through but we hope you find the review thought-provoking. Meanwhile do let us know if there is a topic of interest that you think is worth reviewing in a future monthly letter.
This month, we’d like to discuss the topic of inflation, which once again is a central focus of debate among market participants. Lots of very smart people think we’ll be getting very little in the coming years, and equal numbers of very smart people think we may get a lot. That should be sufficient evidence to conclude that we can’t be sure of the outcome, but the topic is important nonetheless, and getting it right early can be very helpful for performance. Many of the inflation bears (which I will define to mean those who think we won’t get any) are happy to invest in government bonds and tech stocks, confident of a resumption in the, until recently, positive trend in those assets. Many of the inflation bulls are sitting tight with their energy and commodity plays in the expectation that the wind will remain at their back as it has since the year began. So as you can see, it does matter.
Let’s begin with some important definitions. Economists usually use the word inflation to mean increases in the money supply, whereas in common parlance inflation is usually taken to refer to an increase in the general level of prices. Therefore, let us define the former as monetary inflation and the latter as price inflation, to keep the distinction clear. That we are getting monetary inflation seems beyond dispute, but whether we are getting now or will soon get price inflation is a matter of vigorous debate.
With that, we can briefly summarize the arguments made by both sides, and then examine their assumptions and logic.
The price inflation bulls cite the idea that prices are nothing but ratios or exchange rates in which so many units of a given good or service are exchangeable for so many units of money, and so if the quantity of money rises as central banks are causing it to do, so should prices eventually rise, as the new money works its way into the economic system.
They acknowledge that post 2008, when much money was created, it seems we did not get price inflation, and they explain this variously by pointing out that (1) the money was created in a manner that prevented it getting into the economy, and the money is being created in a different manner today, which will lead to price inflation, (2) that the scale of money creation is greater and will have more effect now, (3) that the supply of goods has been restricted by the pandemic and this will continue, which is positive for prices, and (4) that in fact we already are experiencing inflation, but it’s just not being captured in the government statistics
The price inflation bears argue that (1) if we have seen no inflation with all the money printing post 2008, we will see none now, (2) that the economy is weak and the level of debt in the system is gigantic, and this combination is highly deflationary, and (3) that there is spare capacity of labour and manufacturing everywhere, which makes it hard to raise prices.
Now it may be difficult to disentangle all of this, but technically both sides cannot be right, so let’s have a closer look at these issues and see what we can determine.
In classical economics, there was a notion, known as Say’s Law, that supply is what creates demand. In other words, one increased one’s own production in order to be able to afford the production of others. So our production created demand for others’ production.
In the modern era, governments try instead to create demand by simply increasing the quantity of money in circulation, believing that as people come into possession of this money, they will spend it and this increase in spending will raise prices and create demand for new production, thereby goosing the economy into a virtuous circle of activity.
And so they’ve been busy printing money, as shown in the chart below of base money in the United States. Base money represents cash created by the Fed, and it has two components: (1) actual cash in circulation and in bank vaults, and (2) commercial bank reserves at the Fed. The Fed requires banks to hold cash on hand to accommodate depositors who wish to make withdrawals, and they are allowed to do so in one of two ways. Either they can hold cash in a vault on the bank premises, or they can maintain a reserve account at the Fed, which the Fed will convert into cash on demand by the bank. Base money is therefore the sum of cash held by the public and cash held by the banks in their vaults and at the Fed.
It’s very clear from the chart that the quantity of base money increased dramatically after 2008. But base money is only part of the money supply available for spending. A larger part of the money supply are checking and savings accounts, and money market funds, which individuals and businesses hold as balances with commercial banks and brokers, as shown in the chart below. These money components, along with currency in circulation, aggregate to a monetary measure known as M2. As you can see, this fuller measure of money at around $19tn currently is much larger than the base money, which the central banks create directly.
And it too grew quickly after 2008, but not nearly as quickly as base money did, as can be seen in the chart below. While base money rocketed higher, M2 growth grew only as fast as around 10% in 2009 during the recession, and again shortly after in 2012.
So why did base money rise so quickly, but the broader money supply rise so much more slowly?
And now we need to take a moment to review how modern money comes into existence. Central banks create base money out of thin air simply by buying things (usually government bonds) from commercial banks, and they pay by simply dialling up the commercial banks’ account balances at the central bank; that dialling process creates new money - base money. Commercial banks also create money out of thin air in a similar fashion. When you walk into your bank and are granted a loan, the bank simply dials up your account balance. Before the request, that money did not exist. Of course, when you spend that money, it may wind up in an account at a different bank, but it represents an increase in money in the overall system. Corporations also can create money by issuing short term debt, which is acquired by money market funds. All of these things - cash, bank balances, and money market balances - are money that you can spend more or less directly, but they are clearly somewhat different from one another. The components of money that are created by banks and corporations are generally referred to as credit.
So the reason that broad money did not grow as quickly as base money is that credit did not expand as quickly as base money. Before 2008, the Fed allowed banks to create credit in aggregate, only to the limit of 10 times the amount of their reserves held at the Fed, which is effectively how much cash they have on hand. After 2008, the Fed, in an effort to suppress short term interest rates, created a large increase in base money, which initially wound up in bank reserves, as the Fed bought bonds from the banks. This helps to suppress short term rates, because it creates a surfeit of liquidity at the banks. To maintain that liquidity, and also perhaps to avoid the inflationary effect of too much credit creation, the Fed began to pay interest on these reserves to incentivize the banks to maintain large balances rather than using these reserves to increase credit. The effect reduced the debt service costs of many existing corporate lines of credit, which are floating rate in nature, and supported the financial markets and thereby the economy via the wealth effect, without allowing large credit creation and possible inflationary effects.
So the bond market, the stock market, the property market, and other investment and speculative markets rose, increasing the wealth and spending of generally wealthier people. But wealthier people tend not to spend much of their gains, as they already have most of what they need, and food and everyday necessities are a small part of their budget. The increase in money did not make its way deeper into the economy, where the average person was more likely to spend it and thereby directly move the economy.
But even so, broad money was growing, as the chart below of M2 shows, at between 5% and 10% for much of the post 2008 period, so why did inflation only rise at 2%?
One explanation is a weakness in M2 itself as a measure of money. Not every kind of money included in M2 is equivalent to one another. Some money, like demand deposits, can be spent simply by pulling out your debit card, others, like money market funds, are more removed from easy spending. And yet they are treated equally in the monetary averages, such as M2, a dollar of one counts the same as a dollar of the other. Also, certain kinds of money are less attractive to spend than others. If you are earning 3% on your savings account balance, you may be less willing to spend it than another form of money available to you. Also, arguably anything in your brokerage account can be thought of as money, in that you can convert it to money, it’s just that it entails an opportunity cost (because once sold, it no longer can earn a return) and a delay before becoming available. Also, what about credit cards? They are in effect money available to be spent, although they are liabilities rather than assets, so somewhat different that other forms of money. Arguably, in the age of online shopping, a credit card is more useful than cash itself, which is included in M2, while credit cards are not.
In light of these considerations, the Center for Financial Stability created a set of monetary averages, the broadest of which they call Divisia M4, which address many of the issues just mentioned. It was developed out of the work of former Fed economist William Barnett.
Compare the chart of M2 above with the chart of Divisia M4 below, taken from the Center’s website. While M2 growth fell to the very low single digits in 2010, Divisia M4 actually went negative. And in the years that followed, while M2 grew at between 5% and 10%, Divisia M4 grew at only between 0% and 5%.
If Divisia M4 is a better representation of money, then it would explain a lot since the 0% to 5% growth range is more representative of actual CPI growth of 2%, then we would get by looking at M2.
So at this point, perhaps we can answer the question we started with: yes, money did grow after 2008, but not really very quickly. In fact, at a rate not too different from subsequently measured inflation. Perhaps the Fed was too timid in the way they created base money and disincentivized credit creation, to address the economic problems of the time.
But have a look at the growth of Divisia M4 in the past year (and for that matter even M2). If this is a good measure of money, and if more money, in the absence of more real unit output growth, is likely to lead to inflation, then we appear to have some serious inflation coming our way.
As the longer term chart of Divisia M4 below shows, we are in uncharted territory here. Even in the late ‘60s and ‘70s, money growth was much less than today, and we certainly did see inflation back then, not far off what Divisia M4 money growth would have predicted, in fact.
But looking more closely at that chart, you might say hold on a minute, that all sounds fine, but what about the money growth I can see in the first decade of this century? Back then, Divisia M4 grew at between 5% and 10%, and we did not see such levels of inflation following that period.
Which takes us to another important topic to be reviewed in this debate: have we really had so little inflation as the statisticians at the Bureau of Labor Statistics (BLS) who compile the Consumer Price Index (CPI) have been telling us? Or will you believe your lying eyes, which suggest that inflation has actually been a lot higher than 2%?
And because you asked, we now have the opportunity to delve into another important feature of inflation - exactly how it is measured. It is too big a topic to get into the weeds here, but the fact is that the way it is measured has actually changed dramatically since around 1980. Before then, official inflation was designed to measure the change in the cost of a fixed basket of goods and services. Today it is more designed to measure any increase in the cost of living, which sounds the same, but isn’t.
There are a number of changes worth pointing out, and well documented at Shadowstats.com by economist John Williams. The first type of change involves so-called hedonic adjustments. These are quality adjustments. For example, if the computer you bought this year had a faster CPU and more memory than the one made last year, even though the price went up, net of the “improvements”, the Bureau of Labor Statistics (BLS) may determine that the price actually fell. No matter that you need a faster computer to run this year’s software. And note that a drop in quality is not so easily accounted for and therefore ignored. Another change involves so-called substitution. If the BLS noted that chicken fell in price relative to beef, they might conclude that consumers will consequently buy more chicken than beef this year, and so they will raise the weighting of chicken in the index relative to beef to make the basket better reflect consumers’ actual choices. But of course, systematically reducing the weight of goods whose price has risen will result in a reduced measure of inflation, and is no measure of the actual change in the cost of goods and services. And there have been numerous other tweaks designed to systematically suppress the measurement of price increases.
Thus, it is possible that the effect of the rising tide of money earlier this century on prices was offset by countervailing actions by the statisticians at the BLS. In fact, Shadowstats produces a chart of the CPI back to 1980 calculated to show what it would have been if they had not changed the way they measure inflation. This is shown as the blue line in the chart below, as compared to the red line showing measured inflation.
Notice that price inflation as measured by the blue line is more consistent with contemporaneous monetary inflation as measured by Divisia M4.
Also, note that independent measures of actual price inflation show higher rates than the BLS measure. The Chapwood Index attempts to measure the change in prices by looking at the prices each month of the 500 most commonly purchased goods and services in a number of cities around the United States. As shown in the table below, for some years now, price inflation has been consistent with the Shadowstats measure, at around 10%.
So we’ve discussed the development of monetary inflation, how it can be expected to lead to price inflation, how measuring monetary inflation is tricky, and how the way inflation is measured can throw you off. But there is still the mechanism of how money leads to inflation to be discussed, and we don’t want to leave this topic out because it is the subject of much debate at the moment.
So at this point, we need to introduce the so-called velocity of money. Consider that when you buy a sandwich, the seller of the sandwich can take the same money and buy a beer. Velocity measures just how much that money has circulated during the course of the year. It is calculated for the economy in aggregate by taking nominal GDP and dividing it by the quantity of money. So if, for example, nominal GDP is around $21tn and the money supply is around $19tn, velocity would come out to around 1.1. As the chart below shows, it’s been falling for the past 20 years.
There are a host of simplicities embedded in this measure. For starters, GDP does not measure all of the activity in the economy, and perhaps more importantly, M2 is not the only or the full measure of what money is, as we’ve discussed. But the idea is simple enough: if the money velocity is 1.1, for example, it suggests that every dollar in the economy is circulating on average 1.1 times during the year.
Macro-oriented economists will tell you that as velocity has fallen, it has suppressed inflation despite the rising quantity of money. Which only begs the question, why has it been falling? They will then go on to tell you that it has fallen because the quantitative easing of the post 2008 years was executed via the banks in a manner that led to increased bank reserves, rather than showing up in the economy and getting spent. Some of them believe that this will all be remedied, as governments worldwide have learned during the pandemic to increase money by directly providing it to consumers. And when this happens, they expect inflation to come roaring back. That sounds plausible. To say it differently, if the government dropped sackfuls of cash around every neighbourhood in the country, it is not unreasonable to suppose a lot of it would be spent. And since there was no increase in production at the same time, this could be expected to lead to inflation.
But this is not universally accepted as proof of anything. Classically-oriented, and especially so-called Austrian economists, will tell you that the velocity of money is merely a term defined by an economic identity, and it has no use as an explanation for anything. If you want to know why prices rise or fall, you have to look more closely at the incentives and actions of individuals. Those incentives and actions will be driven by such things as their confidence in the future and the consequent need to hold money in reserve, their confidence in the central bank to maintain the purchasing power of the money itself, the level of interest rates, which makes saving more or less attractive versus consuming, and their personal preferences between consumption today and in the future, given their current circumstances. To return to the earlier example of the government dropping sackfuls of money everywhere, we said it would probably get spent, but not if consumers were fearful of the future, then they might feel they need to hold on to the extra money.
Velocity is a ratio which has fallen in recent years because its denominator (the quantity of money) has risen dramatically, while its numerator (the size of the economy) has risen more slowly. Now the nominal growth of the economy (the numerator) itself has two components, real growth arising from increases in population and productivity, and on top of that, the growth relating simply to an increase in prices, i.e. price inflation. If you believe that monetary inflation will eventually lead to price inflation, then the velocity numerator (nominal GDP) should rise, albeit not necessarily at the same rate as monetary inflation, since the numerator has the component of real economic growth to factor in. Recently, real growth and inflation, as measured officially, have been of similar magnitude (prior to the pandemic), so you might expect a fall in velocity when the real economy is weak.
The short answer is yes.
As discussed, the post-2008 quantitative easing was undertaken in a manner designed to support corporates with lower interest rates and to support the private economy generally via the wealth effect by supporting financial markets. They certainly behaved as if they were trying to keep inflation from running away. And as we saw from the growth in money and the rate of inflation as measured by the CPI, it worked.
In the pandemic era, having largely shut down the economy, they clearly felt obliged to support it in a bigger way, and so in addition to plenty of QE, they started handing money out to individuals and businesses - what is usually called helicopter money. Not only that, they increased the scale of money creation quite dramatically, as we saw in the charts above.
This combination is more likely to lead to inflation. While previously they kind of put the brake on credit creation, now they are encouraging it. Many of the pandemic programs in fact use the banks to hand out money by guaranteeing their loans to individuals and businesses. As economist Russel Napier has pointed out, this is likely to become habit-forming. When a politician guarantees a loan, he is not actually increasing government debt, it’s an off balance sheet liability. So the politician can hand out goodies at what appears to be no cost. Of course, we taxpayers are on the hook, and one day, if the economy weakens further, we will get hit with the bill. In fact, most likely they will just forgive the loan altogether.
What many market analysts are beginning to agree on is that the increasing share of GDP that governments have assumed during the pandemic (the US is up from around 20% before to above 30% now, and Europe was above 40% before and now above 50%, and much higher in some countries) is likely to just keep increasing.
At this point, we have just one last topic, which was mentioned earlier as being part of the great inflation debate: what about the deflationary effects of all the debt in the system already, on top of a potentially weak economy?
What we have witnessed since the end of the Bretton Woods gold exchange system in 1971 is an ever-increasing amount of debt. Unfortunately, it is a natural consequence of fiat money and the ability of governments to bridge the gap between what they want to spend and the degree to which people are willing to be taxed by printing up new money and thereby debasing the previously existing money via price inflation.
Since modern money is debt, in other words, it is created by buying government bonds, which thereby back the issuance of the money, then as governments issue money, the amount of debt increases forever. It doesn’t have to be this way, and we have seen countermoves as with the Volcker Fed in 1980, but as the amount of debt reaches higher levels as it has since governments took it upon themselves to fix every economic problem, rather than letting the private sector do so, as it had in the past, the options from this point have become reduced. With as much debt as they have created during the pandemic, the funding required to service this debt at normal interest rates can cripple the economy if not carefully managed.
Consider that before the pandemic, the US collected around $3tn in taxes, but spent $4tn, running a $1tn deficit, in a strong economy mind you. Of that $4tn budget, around 10% or $400bn was interest on the national debt. Post the recent $1.9tn spending bonanza by the US government, the national debt will be on the order of $30tn. If we got a little more inflation and savers demanded an interest rate of 5%, the interest facing the US government would rise to some $1.5tn on that $30tn of debt - about 50% of taxes collected. Such a situation would cause an immediate crisis.
So what will they do?
It certainly appears that they are trapped. They have created so much debt that they can no longer allow interest rates to rise, which means they will have to create even more debt to suppress rates. What can be the end game? On-purpose inflation is really their only way out. If they can allow inflation to rise more quickly, while keeping interest rates low so the economy can continue to grow, the ratio of debt to GDP may fall to a once again manageable level. This is what they did after debts rose during WWII. So-called yield curve control, as discussed in last month’s letter.
In short, they have a serious incentive to allow a controlled level of inflation into the system at this point.
Hopefully, we now have enough information to settle the debate. Ok not settle it, but take a position on the debate.
The price inflation bulls say that
The price inflation bears say that
It certainly looks to us like we will soon see inflation.
How much? The chart below shows the cumulative growth of nominal GDP and money (as imperfectly measured by M3) over time. As you can see, there has been little difference between them. As mentioned earlier, the growth in nominal GDP is mostly a consequence of the growth of money, and since money has very recently grown at a rate rarely seen before, so may nominal GDP surprise on the upside in the coming period.
Bear in mind, we are talking about money growth in the range of 25% (M2) to 30% (Divisia M4). If nominal GDP growth were even half of that, say 12%, and assuming real GDP growth was as much as 6%, which some are forecasting for this year, that would still leave inflation at 6%, high not only by recent standards, but even by historical standards.
We are not talking here about hyperinflation. Hyperinflation is generally agreed to mean inflation at a rate of 50% per month. To get to that kind of rate, money growth would have to be vastly higher, even than it has grown at during the pandemic. But it doesn’t take hyperinflation to wipe you out. If you get inflation of 6% for a decade, that would erase 80% of the purchasing power of the dollar. Any hyperinflation after that would only wipe out the remaining 20%. Six percent is plenty.
And the government is pretty good at creating inflation. Consider that all of the technological innovation we’ve recently been living through is actually deflationary - in that the cost to produce things is always falling. The value of money, if not messed with, should be rising, not falling. Things should be getting cheaper. And yet even using the government's own flawed statistics, the purchasing power of a dollar by their measure has fallen by one third just since the first day of this century, just over 20 years ago.
Well, congratulations if you managed to get this far! We did go on a bit, but we do believe inflation is in our future and will change everything if it comes. The world you are used to of endlessly searching for a few extra basis points of yield will disappear, to be replaced by a panic to maintain purchasing power, and the investments that benefit in that environment could be very different from yesterday's winners.