No Responsibility without Authority: Place Control of Monetary Policy in the Executive Branch?

One of the worst courses I took in undergraduate school was an introduction to management “science”. The only thing I remember from that class is that if one has responsibility for some outcome, one should have the authority to influence that outcome. As we approach the midterm elections, the Republicans are holding the Democrats responsible for the higher inflation that has been experienced in the past two years. The Republicans claim that the high inflation is primarily due to the rapid growth in federal government spending implemented by President Biden and his fellow congressional Democrats, forgetting that a sizeable increase in government spending occurred on the watch of President Biden’s Republican predecessor. Perhaps President Nixon was correct when he exclaimed that we are all Keynesians now.

I will argue, however, that Federal Reserve monetary policy is primarily responsible for the behavior of inflation, not federal government spending. If this is true and the executive and its political party are going to be held responsible for the rate of inflation, then it follows from Management 101 that the executive branch should have the authority to manage monetary policy, not the semi-autonomous Federal Reserve. My wife did not coin the term “econtrarian” to describe me for nothing.

Let’s go through the verbal argument as to why monetary policy, not fiscal policy has the primary influence on the behavior of the inflation rate. If the government increases spending, from where does the funding of this new spending come? It comes from either an increase in taxes or an increase in securities issuance, i.e., borrowing. If taxes are increased, some entities, on net, must decrease their current spending. (There is the possibility that entities could deplete their cash holdings to make their increased tax payments, but generally this would be de minimis.) So, an increase in taxes to fund an increase in government spending would result in approximately a net zero increase in aggregate spending in the economy. The government would spend more; the private sector would spend less. Thus, a tax-financed increase in government spending would not lead to a net increase in aggregate demand, which, in turn, would not put upward pressure on the rate of inflation. The same result would obtain if the increase in government spending were financed by an increase in securities issuance to the nonbank public. To pay for the new issues of government debt purchased, just as was the case for increased taxes due, the nonbank public would have to cut back on their current spending. Thus, a securities-financed increase in government spending would not lead to a net increase in aggregate demand, which, in turn, would not put upward pressure on the rate of inflation. If foreign entities were to purchase the increased amount of government securities, they would have to cut back on their purchases of US exports because there has been no increase in their US dollar holdings. Foreign entities would purchase more US bonds and fewer US-produced Buicks. In sum, there is no logical reason to expect a tax-financed or a securities-financed increase in government spending to lead to a net increase in aggregate spending or an increase in the inflation rate. (Perhaps I am the lone non-Keynesian left.)

Let’s look at some data. Plotted is Chart 1 are the year-over-year percent changes in the annual average of calendar year nominal U.S. government expenditures and the year-over-year percent changes in the annual average All-Items CPI from 1960 through 2021. The growth in U.S. government expenditures is lagged by two years because this yielded the highest correlation coefficient between the two series, 0.51 (shown in the little box in the upper left corner of the chart). Lagging changes in government spending by two years means that the inflation rate this year is associated with the change in government spending two years prior. Remember, if the two series moved in perfect tandem, the correlation coefficient would have a value of 1.00. So, there does appear to be some positive relationship between changes in government spending and the inflation rate.

Now, let’s look at the inflationary implications of increased government spending funded by the Federal Reserve. It is against the law for the Fed to purchase securities directly from the Treasury. But if the Fed purchased an amount of securities in the open market equal to the amount of new government spending, the end result would be the same. So, for pedagogical reasons, let’s cut out the intermediaries and assume that the Fed can and does purchase new securities issued by the Treasury to finance its increased spending directly from the Treasury. The Fed pays for its securities purchase by touching a few keys on a computer keyboard. On the Fed’s balance sheet, the asset item “securities” increases and the liability item “Treasury deposits” increases by the same amount. When the Treasury actually increases its spending, the liability item on the Fed’s balance sheet, “Treasury deposits” decreases. The recipients of the Treasury’s increased spending see their bank deposits increase, which is a liability to the banking system, and their depository institutions (hereafter, banks) see an asset item “reserves at the Fed” increase by the same amount. So far, the Treasury has increased its spending and no other entity has decreased their spending. So, there has been a net increase in aggregate spending, which, all else the same, would impart upward pressure on the inflation rate. But wait, there could be more. The banking system has experienced an increase in reserves at the Fed. If the banking system does not want to hold all of these new reserves, it can expand its earning assets, investment securities and, if there is demand for them, loans until the extra reserves in the banking system are desired to be held. (Presumably, as banks’ earning assets and, thus, deposits, increase, they would want to hold more reserves to protect them against an adverse clearing situation if more of their deposits are withdrawn than new deposits are incoming.) Let’s say that the banking system does increase its loans. Entities typically do not borrow funds to just hold them. Rather they borrow funds to increase their current spending. This increase in current spending by the recipients of new bank loans does not require any other entity to reduce their current spending. So, in addition to the increase in government spending, there also will be an increase in private spending when the banking system converts the extra reserves created by the Fed to fund the increase in government spending into loans and investments. Hence, there will be even more upward pressure on the inflation rate.

Now let’s look at some data comparing the sum of the monetary base (reserves held by banks at the Fed plus currency in circulation) and bank credit (investment securities and loans) with the inflation rate. Let’s call the sum of the monetary base and deposit institution credit, “thin-air” credit because both of them represent credit created figuratively out of thin air. Plotted in Chart 2 are the year-over-year percentage changes in the annual averages of these two series, again from 1960 through 2021. And, again, the highest correlation coefficient between the two series was obtained by lagging changes in thin-air credit by two years. The correlation coefficient in this case is 0.56 compared with 0.51 when the two-year lagged change in government spending was compared with the inflation rate.

Hmm. We have a bit of a conundrum here. Both lagged changes in government spending and lagged changes in thin-air credit seem to have a positive relationship with the inflation rate. It is conceivable that both of these variables have a meaningful effect on the inflation rate. Or it could be that only one of them, for example, changes in thin-air credit have a meaningful effect on the inflation rate and changes in government spending are just “picking up” some of the effect of thin-air credit. In fact, as shown in Chart 3, the percent changes in thin-air credit and government spending are correlated to the tune of 0.41 (the correlation coefficient).

There is a statistical technique that can help disentangle these “cross” effects. It is called a multivariate linear regression. The results of such a regression will yield statistics telling us the probability of whether each independent (or explanatory) variable by itself has an effect on the dependent variable, in this case the inflation rate. Basically, the relationship between one of the explanatory variables is examined, holding the other explanatory variables constant. For example, the relationship between the percent changes in government spending and the inflation rate is established, holding percent changes in thin-air credit constant. Then, the relationship between percent changes in thin-air credit and the inflation rate is established, holding percent changes in government spending constant. By doing this, we can obtain the “pure” effects of each explanatory variable on the dependent variable, the rate of inflation.

But before we get into the results of this regression, let’s look at the relationship between the CPI inflation rate in a given year and inflation rates in prior years. This is what is shown in Chart 4. The highest correlation between the CPI inflation rate in a given year and lagged yearly inflation rates occurs with a one-year lag. And that correlation coefficient is a whopping high 0.81. So, if you were forecasting next year’s inflation rate, you would likely be close to the mark by forecasting that it will be the same as this year’s. In fact, I suspect that this is the “forecasting” model used by most economists, including those at the Fed, and respondents to consumer surveys. It’s a terrific forecasting “model” except when there is the greatest chance to make or lose a lot of money in the financial markets – at turning points. But at least your error will be with the crowd’s error at these inflection points. But I digress. This high correlation coefficient of 0.81 between a given year’s inflation rate and the inflation rate lagged one year suggests that the one-year lagged inflation rate also should be included in the regression as an explanatory variable along the two-year lagged values of the change in thin-air credit and the change in government expenditures.

Let’s end the suspense and report the findings from the multivariate regression. The regression results show that the variables that are statistically significant at a probability of 95% in “explaining” the behavior of annual average CPI inflation rates are, drumroll please, percent changes in thin-air credit lagged two years and the CPI inflation rate lagged one year. The two-year lagged changes in U.S. government spending were deemed to have a probability of 5% or less in having any effect on the inflation rate. That is, changes in government spending do not appear to have any relationship with changes in the inflation rate. (If you would like to see a printout of the regression results, email me and I will provide them to you.)

So, both logic and analysis of the data suggest that percent changes in government spending in and of themselves do not affect the inflation rate. Rather, it is the behavior of a variable that the Fed could control but chooses not to, percent changes in thin-air credit, that has a significant effect on the rate of inflation. If the president is going to be held responsible for the behavior of inflation, then the president ought to have the authority to direct monetary policy. In 2020, the Fed promoted a surge in the growth of thin-air credit to a post-WWII record high of 15-1/2 percent in the face of the COVID-19 negative supply shock. Then the Fed was shocked that inflation rate started increasing in 2021. Growth in thin-air credit was allowed to remain at an elevated rate of 9.7 percent in 2021 compared with a 1960-2021 median growth rate of 7.9%. The rapid growth in thin-air credit then interacted with the negative supply shocks of the February 2022 Russian invasion resulting in a further surge in the consumer inflation rate this year. As the midterm elections are upon us, the president and his political party are being held responsible for the prior incompetence of the Federal Reserve, over which the president has no authority to manage. It is hard for me to imagine that monetary policy could have been worse if the president had been managing it. After growing at an annualized pace of 13.7 percent in the 26 weeks ended December 29, 2021, growth in the sum of the monetary base and commercial bank credit has slowed to 1.7 percent in the 26 weeks ended October 26, 2022. This rollercoaster behavior of the growth in thin-air credit translates into the rollercoaster behavior of the growth, first in real economic activity, then in nominal economic activity. But, whoever occupies the Oval Office is held responsible for it.

If you are concerned about the monetary mischief the executive branch could cause if the Fed were under its direct control, then have Congress pass legislation amending the Federal Reserve Act to require the Fed to hit a specified rate of growth in thin-air credit every quarter. As mentioned above, the 1960-2021 median annual growth in thin-air credit has been 7.9 percent. Over the same period the median annual growth in real GDP and the GDP chain-price index were 3.0 percent and 2.4 percent, respectively. What’s not to like about that? A steady annualized growth rate of 7.9 percent, intuitively, seems a bit too high to me. But if the Fed were required to maintain a steady annualized growth rate of 7.9 percent in thin-air credit, it is highly likely that the amplitudes of business cycles would be damped as well as cumulative increases/decreases in the inflation rate. I’m sure there would be plenty of other things to blame the executive branch for – things over which it does have direct authority.