This title is meant to provoke. The words were first made famous by Milton Friedman, the acclaimed monetarist who believed in a mechanical relationship between money supply and the price level. Such relationship worked pre-Volcker but has failed miserably ever since, at least in advanced economies. Here, I will argue that Friedman was right for the wrong reasons as he failed to recognize the most important inflation driver, namely the demand for reserves.
MMT has brought new prominence to the question of inflation. According to MMT, governments that issue their own currencies do not face financial constraints other than inflation, but even that constraint is not binding at all times. Only at full employment, can inflation become an issue as money-financed government spending starts bumping against limited real resources.
Such line of thinking is misguided as there is no causal link between inflation and employment. Instead, inflation is always and everywhere a monetary phenomenon. Resource constraints simply determine how sensitive the price level is to the underlying monetary factors. At full employment, the sensitivity is high; at low employment, the sensitivity is low.
To illustrate, let’s first take a look at historical data on inflation and the unemployment rate (the unemployment rate being a proxy for resource constraints).
The graphs above show the Phillips Curve in the US for three periods. First, you will notice that the R-squares are super low, suggesting very weak relationship across the entire time horizon. Furthermore, the first period shows an upward-sloping curve, which is the opposite of what one would expect if resource constraints gave rise to inflation. This period includes the stagflation of the 1970s, during which high inflation was accompanied by high unemployment. An upward-sloping Phillips Curve is not unique to the 1970s. To this day, emerging and developing markets face upward-sloping Phillips Curves that subject their economies to the double scourge of high inflation and high unemployment. This means that the underlying cause is neither limited resources, nor oil shocks, nor unions. Something more fundamental must be at play that gives rise to inflation, even during periods with plenty of idle resources.
The second graph roughly coincides with the hey-days of inflation targeting. The downward-sloping Phillips curve is consistent with the resource-constraint hypothesis. However, you will notice that the curve never crosses into negative territory. Only in 2009 did prices barely decline in the aftermath of the Great Recession. There were two prior recessions during which prices rose, despite high unemployment, albeit at a lower pace. The original Phillips Curve was derived from data during the Gold Standard period, and it clearly crossed into negative territory. Back then, prices were steady in the long-term but were subject to wild swings in the short-term. High inflation during booms was quickly followed by severe deflation during busts. The lack of deflation under inflation targeting, even during recessions, means that the resource-constraint hypothesis fails once again.
Last but not least, the lowflation in the years since 2010 should dispel any remaining doubt as to the lack of causality between resource constraints and inflation. The Phillips Curve is flat. During the period, unemployment declined from almost 10% down to 3.9%, but inflation has not budged! Let’s add this to the pile of unexplained macro puzzles about inflation.
To understand inflation, the very first question you need to ask is What is the reserve currency? Reserve here refers to a long-term store of value with zero duration (meaning zero nominal price volatility) and zero credit risk. Governments do not determine what currency can act as a reserve, people do. The supply and demand for reserves are the two most important drivers of the business cycle (I outline this view in a series of recent posts: Alternative View of the Business Cycle).
The demand for reserves, which I refer to as asset money demand, measures the long-term demand for money by agents with negative income-growth expectations. To visualize asset money demand, think of a person who believes her stock portfolio will decline in value. Accordingly, she sells some of her stocks, allocating a larger portion of her long-term savings into money. That demand for money is asset money demand (Keynes referred to it as speculative demand for money).
In advanced economies, central bank independence and commitment to price stability have established the domestic currencies ($, €, £, ¥) as reserves capable of satisfying asset money demand. That is not the case in many emerging and developing economies, e.g., Turkey, Brazil, Russia, etc. Domestic agents there prefer foreign currency for their asset money demand, giving rise to a positively-sloped Phillips Curve. It is important to note that commodities such as gold or oil can also act as reserves, the Gold Standard being the go-to example.
Before diving into the question of inflation, let’s first discuss the most fundamental feature of a monetary economy, namely that demand comes first and creates its own nominal supply. Here I reprint a chart from the post I referenced above.
In a monetary economy, before any exchange of goods and services can occur, money must be created first. Banks create money by extending credit to borrowers. In other words, the private sector spends money into existence by borrowing from banks (think of an economy where everyone uses credit cards to facilitate transactions)[1]. Furthermore, we borrow in anticipation of future incomes. This means that a monetary economy bootstraps, whereby incomes must be anticipated ex-ante to be generated ex-post [2]. Banks are critical since they are the flux-capacitors that enable the economy to bootstrap. Banks finance aggregate demand by creating money. As agents spend the money, they generate nominal incomes, thereby fulfilling their incomes expectations. Agents use such incomes to pay-off the debts incurred at the beginning of the period (eg, paying-off credit cards). Upon such pay-off, the corresponding money balances are destroyed (credits cancel debits).
Aggregate demand is given by income expectations for the current period plus desired borrowings less desired savings. For a moment, let’s assume that desired borrowings equal desired savings. This means that aggregate demand equals income expectations at the beginning of the period. Income expectations are given by the expected nominal aggregate supply (expected supply multiplied by the expected price level). As a result, aggregate demand equals expected nominal aggregate supply. In other words, when desired borrowings equal desired savings, the economy produces and consumes exactly what is expected, hence actual price level matches the expected price level. For purposes of illustrations, let’s assume that the expected price level is consistent with actual prices observed in the prior period. This means that the price level remains unchanged. Relative prices may change during the period based on shifts in preferences or production shocks affecting individual goods, but the aggregate price level will remain the same [3].
The most important take away here is that, if desired borrowings equal desired savings, the price level will be unchanged regardless of any resource constraints. The economy can enjoy steady prices even at full employment as long as all expected incomes are fully spent, no more nor less. On the flip side, the economy can be stuck in a steady state of high unemployment. Unless someone bootstraps aggregate demand for the potential incomes of unemployed agents, aggregate demand will be insufficient to provide them with jobs.
In a Walrasian economy, prices equilibrate demand with supply, meaning that if workers accepted lower wages, entrepreneurs will provide them with jobs. In a monetary economy, demand is determined ex-ante, and prices are simply one of the unknown residuals. Asking for lower wages, simply shifts relative demands for labor. You do not make the pie bigger by asking for a smaller slice. Nor by asking for a larger slice, which goes to show the absurdity of the claim that worker wage demands cause inflation. Wages simply distribute ex-post incomes.
Going back to the borrowings and savings scenarios, if desired borrowings exceed desired savings, you have aggregate demand in excess of income expectations. Assuming an upwardly-sloped aggregate supply curve, firms respond to such excess by increasing both prices and employment. If desired borrowings are less than desired savings, you have shortage of aggregate demand (or demand below income expectations for the current period). Firms respond by cutting both prices and employment. It is precisely this dual response to either shortage or excess of aggregate demand that gives rise to the Phillips-Curve relationship. However, the Phillips Curve does not describe causality but coincidence. The causality lies with desired borrowings and savings.
Next, let’s take a closer look at capital and money markets where borrowings and savings trade:
- Capital markets are for assets and liabilities with duration. Duration means that nominal prices, even for risk-free assets and liabilities, are subject to change with changes in interest rates, but expected returns are fixed (in the absence of credit risk). Capital markets do not have a market-maker who guarantees nominal prices. This means that changes in asset prices (aka, interest rates) equilibrate demand with supply. In other words, for every duration borrower, there must be a willing duration saver, and vice versa.
- Money markets are for assets and liabilities with zero duration. Zero duration means that the nominal price of risk-free assets and liabilities (aka money) is not subject to change (even in a changing interest rate environment), but the expected asset returns are variable (think credit card debt or interest-bearing demand deposits). Unlike capital markets, money markets do have a market maker, namely banks. Banks stand ready to facilitate zero-duration borrowers even in the absence of willing savers. On the flipside, banks can accommodate long-term savers, who want to hold their savings in money (think asset money demand), even if there are no long-term borrowers on the other side.
A disequilibrium between desired borrowings and savings can occur only in money markets because money markets have a market maker while capital markets do not. Such disequilibrium gives rise to NGDP fluctuations, including changes to the price level, that equilibrate actual borrowings and savings as well as money supply and demand [4].
Money supply and demand have two components each. The money supply includes the monetary base, supplied by the central bank [5], and endogenous money, representing credit-line and credit-card balances incurred by bootstrapping agents. Endogenous money is a residuals of NGDP because agents use revenue and wages earned during the period to pay-down their credit lines. At the end of the period, whatever is left over equals the endogenous money supplied by banks to finance bootstrapping agents. Money demand includes asset money demand, which I described above, and transaction money demand. Transaction money demand represents the money held by non-bootstrapping agents to facilitate transactions in future periods. Transaction money demand is also a residual of NGDP because agents use their actual incomes to replenish their money balances after the expenditures incurred during the period.
The two components of money supply and demand that are not residual of NGDP, namely the monetary base and asset money demand, can result in disequilibrium between desired borrowings and savings, giving rise to fluctuations in NGDP and changes to the price level.
Let’s assume, that domestic agents consider the domestic currency a reserve capable of satisfying their asset money demand (eg, the US). If the central bank supplies reserves in excess of asset money demand, this results in excess desired borrowings that cause employment and prices to rise. Effectively, there is not a willing saver on the other side of the bonds that the central bank bought. The result is no different if the excess monetary base is supplied through money-financed fiscal deficits. Such government spending is incremental to private activity because there is not a willing saver who chose to forego consumption. The excess supply of base money also causes interest rates in capital markets to fall below income-growth expectations (either because the central bank bought assets with duration directly from capital markets or because money-financed fiscal deficits prompted agents to buy such assets to restore their desired allocation into money). As a result, asset prices rise above fundamentals, which is consistent with the pattern of asset bubbles in economies with inflation-targeting central banks.
The lowflation in the aftermath of the Great Recession is an interesting case study. The housing bust severely depressed income-growth expectations, which led to a massive increase in asset money demand. This explains why QE was non-inflationary. It’s important to note that at rates close to ZLB, bonds and reserves become close substitutes. Accordingly, the Fed cannot oversupply base money by buying bonds. As a result, monetary policy was not in position to bootstrap a recovery by relying on excess borrowing by the private sector. However, through QE, the Fed monetized the fiscal deficits during the period, ensuring that government spending was incremental to private activity. Accordingly, the excess borrowings that bootstrapped the recovery and resulted in low inflation came from the public sector. EU makes an informative parallel. Despite aggressive QE by the ECB, the economies in Europe have not recovered fully due to insufficient fiscal deficits.
The second scenario with respect to asset money demand covers countries where domestic agents prefer foreign currency for their asset money demand (e.g., emerging and developing economies). It is well known that economic conditions in such countries are dictated by foreign capital flows (and by extensions, the foreign currency exchange rate). That’s because domestic agents seek foreign currency for their asset money demand. In recessions, asset money demanded rises since more agents have negative income-growth expectations. This causes domestic capital flight into FX, which leads to a fall in the FX exchange rate. This causes sellers to increase domestic prices (the prices that matter to everyone are the FX-denominated prices, not the domestic ones). As a result, such economies experience both high inflation and high unemployment.
If the central bank or the government were to increase the monetary base either through asset purchases or monetized fiscal deficits, this simply causes the FX exchange rate to fall, resulting in higher inflation. That’s because there is no underlying asset money demand for the domestic currency. As soon as the government prints the new reserves, agents try to exchange them into FX. Governments simply find themselves unable to manage the business cycle unless they borrow FX in international markets.
The mechanical relationship between money supply and prices that Milton Freidman believed in describes precisely a country where asset money demand is not denominated in the domestic currency. The US also went through such period. In the aftermath of the gold exit by the Nixon Administration, domestic agents continued to prefer gold for their asset money demand. This gave rise to the 1970s stagflation. Only after Fed Chairman Paul Volcker took bold action to restore price stability, did the US dollar regain its role as global currency reserve. This severed the mechanical relationship between money supply and prices and provided the policy space for active monetary and fiscal policy ever since.
This discussion goes to show the importance of price stability. Of the two central bank mandates, price stability clearly dominates. Without price stability, the domestic currency is not perceived as a store-of-wealth. As a result, domestic agents seek some other reserve to satisfy their asset money demand. This gives rise to a binding inflation constraint that prevents policy makers from implementing full-employment policies.
The global reserve status of the US dollar is the Volcker dividend we benefit from to this day. We must strive to preserve this dividend (if history is any guide, the reservoir of good will when it comes to money is not deep). With respect to full-employment policies and asset bubbles, we do need to re-examine monetary and fiscal policies and create better tools, as I advocate in this post on the right policy mix. However, such policy re-think must not compromise central bank independence and the commitment to prices stability, the two pillars of any reserve currency.
[1] MMT claims that governments spend money into existence. Under the current legal framework, at least in the United States, that is not correct because the Treasury is prohibited from over-drafting its accounts at the Fed. MMT will counter that the Fed funds spending by the government by injecting reserves prior to each auction of Treasury securities. However, MMT omits the important detail that the primary dealers, who buy the newly-issued securities, duration-match. As soon as they buy the securities, they put on a duration hedge by selling fixed-payer swaps. This means that the bank money, in the hands of the variable-swap payer, must exist prior to the auction. In other words, in coordinating with the Treasury auction, the Fed simply exchanges bank money for reserves. The Fed acts as nothing more than a currency exchange bureau, exchanging bank money for reserves, thereby enabling the private sector to pay taxes and buy government bonds.
[2] Agents in a monetary economy can generally be divided into two groups, bootstrapping entrepreneurs and non-bootstrapping workers. Entrepreneurs borrow from banks to hire workers and produce an inventory of goods and services. The newly-created money flows to workers in the form of wages, who now have the money to buy the very same inventory they themselves produced. The revenue from such sales allows the entrepreneurs to pay-off their bank credit lines. In other words, entrepreneurs bootstrap on behalf of workers, which goes to show the importance and self-fulfilling nature of “animal spirits”.
[3] Are there production shocks that can affect the aggregate supply curve (or the production of all goods) in the sort-run? That’s unlikely because most of the production inputs are not subject to such aggregate short-term shocks. Unless we are dealing with pandemic or total war, the labor input is driven by demographics or long-term tends with respect to the participation rate. Capital is fixed and depreciates at a steady rate. New investments in capital do not affect the existing AS curve but only whether the curve shifts to the left. Technology cannot be unlearned. Land is the only factor that may be subject to short-term shocks; however, in a large economy (i.e., the global economy) such shocks largely cancel out as disasters in one part of the economy are offset by bumper crops elsewhere. How about oil shocks such as the oil embargos in the 1970s? Again, in a bootstrapping economy where incomes must be anticipated ex-ante to be generated ex-post, such supply shocks simply shift relative demands (i.e., shock to the price of oil shifts demand away from, let’s say, the restaurant sector causing real wages of restaurant workers to fall).
[4] According to Keynes, fluctuations in NGDP, not interest rates, enforce the I=S identity. However, Keynes believed that interest rates equilibrate money supply and demand (i.e., the interest rate is price for parting with liquidity). I show that fluctuations in NGDP also enforce the money supply and demand identity, irrespective of interest rates. That’s because endogenous money, created by banks to finance bootstrapping entrepreneurs, and transaction money, demanded by non-bootstrapping workers, are both residuals of NGDP.
[5] Technically, the reserves supplied by the central bank are not part of money supply because central bank reserves are assets on bank balance sheet. For the purposes of this discussion, though, I consider the offsetting bank liabilities, which are part of the money supply, as having been monetized by the central bank.
