Modern Monetary Theory (MMT) has always bothered me in the caviler way it treats inflation. It recognizes the problem but assumes that it is easily managed. Here I present supporting empirical evidence that inflation is feature of the method used to create fiat money.
To find empirical evidence in a working economy, you first need a working definition of money, a source for creating money, a reason for expecting money to be used in the general economy, and then a theoretical link between money and prices. We will build a framework that will (hopefully) accomplish this.
Our framework will describe a system capable of including a nexus found between banks, money, and incremental access into the general economy. In the interest of brevity, we won't spend much time on definitions, depending instead upon the common knowledge of how all of us conduct business to provide most of terminology and associated relationships.
We try to keep it simple.
Most of the money on deposit at banks is owned by people who traded something to acquire it. A fraction of deposits are owned by people who placed borrowed money on deposit. Both cohorts have equal access into the general economy, limited only by the size of their bank accounts.
We could add up all bank deposits and create a yearly report. This number would be unchanging if borrowers only took loans from people who first traded something to acquire money. The deposits number does change so money must be created in some fashion. New money must enter the general economy through some mechanical methodology. We'll put a methodology into our framework.
Definitions and Assumptions
We will define 'fiat money' as being that money used by the national government to pay it's bills [*]. This money must come from a source.
We will define the source of fiat money as being banks, both central and private. Private banks are assumed to operate under a government license so that they can legitimately create real fiat money [**].
We assume that banks create money when they make loans. This is a key assumption which is widely accepted by mainstream economic theory. To be granted a bank loan is quite a privilege .
We assume that most economic entities use fiat money for transactions and that this money is stored in banks during the time it is owned by any entity. Entities like to own money because the ownership of money provides incremental transactional access into the general economy.
This concludes the list of basic assumptions that we need. We will make some procedural assumptions as we move along.
The first procedural assumption relates to who owns the money represented by bank deposits. We will assume that workers own the deposits created when they deposit earnings. In a similar fashion, we assume that sellers of material items own the deposits created as a result of making sales. Borrowers own the deposits created by borrowing from a bank.
We resolve the quandary introduced by the first procedural assumption.
We recognize that a quandary that has been introduced by the combination of our definitions and the first procedural assumption. Bank deposits represent both earned and borrowed money while, at the same time, all fiat money is created by borrowing. We resolve this ambiguity by assuming that a borrower's deposit is composed of newly created fiat money (created under license from government). Once borrowed money is spent, it moves into the ownership of entities who first trade labor or trade material property. You can see that the vast majority of money on deposit is owned by entities that first trade labor or trade material property.
Our assumptions are complete.
We have all the assumptions we need. Nearly all readers should be able to recognize familiar aspects of our assumptions. Next, we need to relate them to empirical data commonly found within macroeconomic analysis venues.
The Empirical Connection
Money supply and bank deposit data are routinely published by government agencies. We will use "Deposits, All Commercial Banks" to build a chart. This data gives us a measure of the money available to entities for making transactions.
It is great to know how much money is available for making transactions but we need to know more if we wish to relate money to the general economy. This task is not so easy to do. One routine report published by government, GDP, tries to do this but has many built in assumptions that seriously decouple the real economy from the money supply available to facilitate transactions. The GDP report tries to relate prices with total transactions, giving us an approximation of how much economic activity occurs within a time period. A simple relationship between total money supply and economic activity reported by GDP can be made (MV = PT) but I would call it somewhat contrived.
A tighter link between money and economic activity can be made if we take advantage of a distinction that we found in bank deposit ownership. We will focus on borrowing owners who would be spending newly created money. Newly created money would expand the measured money supply which would be detectable from periodic measurements as a change in totals. We can (safely?) assume that borrowed money would be rapidly spent in an event captured by a GDP measurement. It follows that changes in periodic GDP would be at least partly caused by a change in money supply, which would be the result of changed within-period borrowing. We will relate changes in GDP to changes in our measurements of money.
There are a lot of economic conditions that can change GDP measurements from year to year. External shocks (such as COVID) and borrowing based spending, can both cause rapid changes in periodic GDP measurements. Slower GDP changes occur due to inflation (meaning a price change without a change in volume), product preference changes, foreign trade, and evolving technology.
What does a graph of the periodic changes in deposits, and GDP look like? Figure 1 displays traces of the periodic changes (expressed as per cent) in "Deposits, All Commercial Banks" and GDP.
Comments on Figure 1
The correspondence of traces in Figure 1 is both expected and surprising. Expected because it is what we theorized. Surprising because the GDP data is so imprecise and time periods between sets of data are likely unequal.
Newly created money is the cause of the increase in deposits. Inflation is likely the principal cause of persistent GDP changes in one direction. The correspondence found tempts us to think that newly created money is the cause of inflation. However, it may be more accurate to say that increases in the supply of money allows inflation to be spread over the economy and become embedded. Further study could be done on this topic.
I am going to refrain from further comments on Figure 1 except to notice that the chart's time span extends through some notable economic disturbances that might be reflected here, or maybe, surprisingly, not reflected:
Not reflected: The 2008-9 massive stimulus (in the bank deposits trace)
Reflected: The 1980's inflation and sharp recession, the 1992-3 dot-com building spree, the 2000 dot-com recession, the 2008-9 recession (in the GDP trace), the current COVID related economic distortions.
This chart is something I will puzzle over for a while. I can't recall seeing anything quite like it before. However, I am far from being a trained economist with a solid historical background. Probably others have seen similar charts, particularly if they are of the Monetarist theory persuasion.
The Framework in Nutshell
1. Government pays by using real fiat money.
2. Within qualifications, banks create new fiat money when they make loans.
3. Entities prefer to keep fiat money in banks because ownership of fiat money stored in banks allows easy access into the general economy. It is convenient to exchange money for labor and property, whereby a unique quantity of money[M] equals unique item of property[P] or [M = P].
Supporting real world evidence can be found by relating the change in the supply of fiat money to changes in the number of transactions and changes in the M = P relationship. The change in money supply can be found by taking advantage of two different methods of establishing a positive bank deposit balance.
Here, at the Mechanical Money blogsite, I've had a long search for a comfortable economic paradigm that fits the economic reality that I have experienced over many years. The assumptions and relationships presented here seem to conform to the real world very well.
The actual bank mechanics of creating and transferring money are not part of the macroeconomic framework presented here. 
The chart in Figure 1 reflects persistent increases in both money supply and GDP. This persistence can be explained by assuming persistent inflation. However, the fact of close association does not by itself establish a cause/effect relationship. However, it is easy to make a case that increased borrowing will stimulate demand for consumer goods that are ready for purchase.
A loan by a bank to a customer is the grant of quite a privilege. This loan establishes an account with a deposit which, from an access into the economy perspective, gives the borrower the same access into the economy as is enjoyed by those who worked or sold material property to gain access.
When new money is delivered to the borrower, borrowing from banks can be describe as taking the form of layers of money supply growing on a periodic basis. Because money is used in commerce, measurements of the amount of money used in commerce can be made. The paradigm incorporated here hinges around the ownership relationships found between banks, money and incremental access into the general economy.
I view economic theories such as Monetarist, Keynesian, and MMT as being mostly political formulations, valuable in a political arena but not mutually consistent in mechanical detail. When paradigms revolve around the Bank-Money-Access nexus, they seem more based in mechanical reality.
 The privilege granted is the ability to borrow from the generalized pool of existing wealth owned by the ongoing private economy. This action is related to a privilege belonging to government known formally as seigniorage. When exercised, both privileges increase the supply of money in the economy beginning at the instant of being recognized by the private sector.
 The concept of banks as financial intermediaries has been massively modified here. The intermediary argument depends upon bank's reuse of customer's money in the sense that banks 'borrow short and lend long' but does not provide for an increase in money supply. We could enter a long discussion of the merits of that approach but the discussion would take our eyes off the essential concept that must be kept in mind: the bank borrower receives a loaded bank account that gives him a spending position identical to the spending position enjoyed by people who work or make sales to achieve the identical spending position.
Some concepts of money have been discarded. Money is not debt of government. It is just a token that represents some measure of incremental access into the general economy. The value of the unique token, whether paper or electronic notation, can only be based on a history of usage of similar tokens. Private banks routinely extract a pledge from the borrower to return the borrowed tokens.
(c) Roger Sparks 2021