Latin American corner: Neo Fisherism, New Keynesianism and monetary policy in Latin America (II)

By Naked Keynes (Anonymous Guest Blogger)

The positive relationship between nominal interest rates and inflation is not a new stylized fact in economic theory. In the 19th Century Thomas Tooke (1774-1858) considered it a general rule illustrated by the data presented in his History of Prices and the State of Circulation, 1792-1856 (published over the period 1838-1857).

Tooke rationalized the positive relationship between inflation and interest rates by postulating that the interest rate is part of the cost of production of commodities As a result when interest rates rise so does the cost of production and hence prices. As he put it in his Inquiry into the Currency Principle (1959 (1844) p.81: “A general reduction in the rate of interest is equivalent to or rather constitutes a diminution in the cost of production…the diminished cost of production hence arising would…inevitably cause a fall of prices of all the articles into the cost of which the interest of money entered as an ingredient.” Tooke dismissed the existence of a negative relation between interest rates and commodity prices. A fall in interest rates may be synonymous with liquidity but as Tooke remarked (Ibid, p.79):“A power of purchase might thus doubtless be created; but why should it be directed to the purchase of commodities if there was nothing in the state of supply, relatively to the rate of consumption, to afford the prospect of gain on the necessary eventual resale?..The error is in supposing the disposition or will to be co-existence with the power. The limit to the motive for the exercise of the power is in the prospect of resale with a profit.” By the way the assumption of equating the disposition with the power of purchase is a fundamental tacit assumption underlying the monetarist, New-Classical and New Keynesian monetary transmission mechanisms. It´s the whole story behind real cash balances and the transactions demand for money. Tooke went further he assimilated the effect on interest rates on asset markets (Ibid, p.86):“A low rate of interest is almost synonymous with a high price of securities…”

Tooke´s views were challenged by Knut Wicksell (1851-1926): “Tooke´s thesis is certainly wrong…The argument is based on the inadmissible, not to say impossible, assumption that wages and rent would at the same time remain constant, whereas in reality a lowering of the rate of interest is equivalent to a raising of the shares of the other factors of production in the product” (Wicksell, Lectures on Political Economy, II, p. 183). Wicksell´s criticism of Tooke and his disciples led him to explain the rise in prices and inflation by the gap between the natural rate of interest (“the rate of interest at which the demand for loan capital and the supply of savings exactly agree”, Ibid, p. 193 and which depends on real as opposed to monetary factors including “the efficiency of production…the available amount of fixed and liquid capital, on the supply of labour and land…” Wicksell, Interest and Prices (1936 (1898) p. 106)) and the money rate of interest. Assuming full employment, a pure credit system and that banks respond endogenously to the demand for credit Wicksell showed that when the natural rate of interest exceeded the money rate of interest an inflationary process ensued. The inflationary process led eventually to an increase in the money rate of interest to match the natural rate of interest at which point inflation would stop. In this way Wicksell was able to resurrect the positive relation between the money rate of interest and inflation while rejecting Tooke´s theses.

The Wicksellian distinction between the natural and money rate of interest is at the heart of the New Keynesian model. It appears in the aggregate demand equation (IS) and in the Taylor rule and in fact the Central Banks that have explicit inflation targeting regimes (as well as some that do not) must obtain estimates of the natural rate for their models to be operative. Without the natural rate there would be no New Keynesian monetary models. However, it is odd, that the relation they postulate between the money rate of interest and inflation is exactly opposite to that of Wicksell.

Needless to say, New Keynesian inflation targeting models do not have scope or space to include the type of causality between interest rate and asset markets envisage by Tooke (which is an essential component of Keynesian economics). These models do not include the banking system or asset markets. This is due to their firm commitment to the upgraded “divine coincidence”: price stability (equating the market and natural rate) is equivalent to full employment and to financial market stability.

To be continued

Lara-Resende, Cochrane and the Brazilian Recession

GDP has collapsed by a bit more than 7% in real terms over the last two years in Brazil (graph below show more recent data). This constitutes the worst crisis in recorded macroeconomic history, worse than the debt crisis of the early 1980s, and even the Great Depression.

The reasons for this crisis are entirely self-inflicted. I discussed those issues before here (and here). The problem is not fiscal, which resulted from the crisis, nor external, since there was no real issue in financing the current account deficits. The fiscal adjustment was the main cause of the recession. And certainly monetary tightening didn't help, actually it made the fiscal situation worse by increasing debt servicing costs. At any rate, recently a short newspaper piece (in Portuguese, and registration might be required) by André Lara-Resende, one of the authors behind the idea of inertial inflation, and a student of Lance Taylor at MIT in the 1970s, has received significant praise from a wide and diverse audience. So I finally decided to read it.

Lara-Resende notes, correctly, that the empirical evidence for the simple Quantity Theory of Money is not particularly good, and that there is little connection between money supply and prices. He also notes that modern macro has had a Wicksellian turn, something that is not that new. He is retooling and trying to learn modern macro at Columbia, and discovered the Fiscal Theory of the Price Level (FTPL).

In the traditional QTM story, if the economy is at full employment, and if we assume that the optimal output level is rigid and does not react to demand expansion, and the central bank increases money supply, then inflation would follow. Inflation may actually follow in this scenario even if the central bank decides to tighten monetary policy and not finance the deficits, according to New Classical authors. Sargent and Wallace argue that economic agents know that lower money supply growth today would lead to faster money growth in the future. In other words, if the central bank does not monetize debt in the present, it will be forced to do so in the future, since public debt is a promise to pay in cash in the future, and tighter money now means more inflation later. This would be a situation in which fiscal policy dominates monetary policy.

Fiscal dominance is a situation where the central bank gives up its control over the quantity of money and over inflation to prevent the government from defaulting on its public debt. In addition, with rational expectations, economic agents would anticipate this higher growth rate in the future, and inflation would accelerate in the present without monetization. Sargent and Wallace suggest that inflation becomes always and everywhere a fiscal phenomenon. This is the basis of the FTPL, and the work of Cochrane, which Lara-Resende sees as unorthodox and an alternative to conservatism.

Note that he sees this ideas as unorthodox, in particular because they can account for the possibility of a positive relation between the nominal rate of interest and the price level, something that used to be called the Gibson Paradox.* But, like Cochrane and his fellow New Classical authors, the idea is that to control inflation you need a good old-fashioned fiscal adjustment (btw, that was true too for the old Monetarists, that thought that monetary expansion was caused by fiscal deficits).

The notion that monetization of debt, or fiscal dominance, would lead to inflation, however, is predicated on the supposition that the system has a tendency to move to full employment, and to do it relatively fast. The central question regarding monetization, then, is what determines this supply constraint that imposes an inflation barrier to demand expansion, what Friedman referred to as the natural rate of unemployment – note that natural was meant to suggest that policy cannot affect it. If the economy is below the supply constraint or the supply constraint is variable, then monetization is not inflationary.

Interestingly enough, Lara-Resende does not discuss whether Brazil is at the natural rate at all, and what would be the effect of a massive fiscal adjustment. The idea here is that low rates of interest would signal, lower inflation expectations, and that the fiscal adjustment would reduce the real rate of interest determined by non-monetary factors (the natural rate of interest). There are too many problems with the notion of a natural rate of interest (yes, as I always insist, go learn about the capital debates). But there is hardly anything unorthodox about it. And besides, austerity, which seems to be the default position now in Brazil, would lead to a worsening of the crisis, and contrary to what Lara-Resende thinks, monetary tightening, through its effect on the exchange rate and the cost of imported goods (and its impact on real wages; on top of the effects of the recession on workers' bargaining power, and the government's policy for minimum wages), would maintain inflation low. Low rates of interest (even if welcome) are not the solution; at this stage only expansionary fiscal policy would work.

* It's not really a paradox in a classical (as in classical political economy) context, in which the interest rate is a cost of production.

Clarida on Fed policy: or how does the Fed affect inflation

Richard Clarida gave an interview (right at the beginning of the podcast) on why the Fed should increase the rate of interest. He also said that the Fed can affect inflation, which, he correctly points out, is denied by several economists. However, the degree of confusion on this subject is significant, and modern monetary theory, and its implications for central banking behavior, is, in part, responsible for that.

The conventional wisdom on what central banks can do (and one can think of Clarida's contribution with Galí and Gentler as a good summary of the dominant position) suggests that central banks can only target inflation efficiently in the absence of cost-push inflation or if they do not care about deviations from the natural rate of unemployment (or output, depending on whether the unemployment gap or the output gap is used for the central bank's policy rule). And some authors believe that inflation is in the long run really always caused by excess demand, and that the system gets back to its natural level, so that there should be no concern other than inflation.

Under these circumstances the Fed has an effect on inflation through two mechanisms.The Fed could maintain the rate of interest low even in the face of an economy close or in extremis at the natural rate, supposedly stimulating spending, and it could also affect inflationary expectations. The later is the basis for Blanchard's proposal to increase the inflation target to 4%, for example. The same solution was suggested by Krugman (my comments here).

On the first issue, besides the problem that most mainstream economists do not know what is the precise level of the natural rate of unemployment (and I'm not even going to get into the logical problems which the rate), there is the issue that the natural rate is not constant, and it might be affected by the actual level of unemployment. So as the Fed keeps the rate low, and again this would arguably lead to more spending, higher output and lower unemployment, the natural rate would fall, and inflation might not pick up.* In addition, there is the fact that lower rates might not translate into higher private spending after all. Eccles famous 'pushing on a string' story. In that case, it seems that the Fed has actually limited capacity to affect inflation.

The second point emphasis the role of expectations. The notion is that the Fed affects what agents think and has real effects through this channel. This is a variation of the inflation expectations fairy according to which higher expected inflation would lead to more investment. Here again Eccles is relevant. Eccles was very explicitly against confidence fairy type arguments (see here and here). Here the argument is that if the Fed announces its willingness to allow for higher inflation, then agents would assume that inflation would indeed be higher (economic agents don't seem to care what the mechanism by which inflation will end up being actually higher), and would spend more (holding to cash would hurt them).

So if the Fed says inflation could go up to 4%, even though oil prices are low, and wages are stagnant, economic agents ignore facts, and hold to the expectations of inflation. In this case, by spending, with the economy being at the natural level, economic agents cause the inflation that the Fed stimulated in a self-fulfilling prophesy. Again, inflation occurs because of excess spending when the economy is at the natural rate. The key to understand how (what mechanism) the Fed can (or can't) affect prices is crucially linked to that very problematic concept of the natural rate.

Don't get me wrong. I think the Fed can affect prices, but is way less effective than what the conventional wisdom suggests, and its powers are asymmetric, meaning it can reduce inflation more than it can increase it. First, hiking the rate of interest, from a cost perspective, increases financial burden of firms and that might lead to higher prices (something that was known as the Gibson Paradox, the positive relation between the interest rate and the price level; later has been also known as the Cavallo-Patman effect). But higher rates, more importantly, increase the burden of debt, leading consumers and firms to possibly default, and reduce the level of activity. It is the lower level of activity, and its impact on workers bargaining power, and, hence, the absence of wage pressures, that matters for stabilization.

On the other hand, reducing rates as noted above might have little impact on spending and the level of activity. Traditionally low rates were instrumental in making fiscal expansion easier, since the government could borrow at low rates. But in the absence of government spending, and any other additional increase in autonomous spending, lower rates per se are unlikely to lead to a booming economy, particularly one that is beyond full capacity and that would display demand-pull inflation (inflation more often than not is cost-push anyway).

* Laurence Ball now says the natural rate of unemployment is around 4.3%, for example.

Thomas Tooke and the Gibson Paradox

The Gibson Paradox is the name that Keynes suggested for a particular empirical regularity, namely: the positive correlation between the rate of interest and the price level. He had read about in a series of articles in the Banker's Magazine by A. H. Gibson, hence the name. Keynes suggested in his Treatise on Money that the Gibson Paradox was "one of the most completely established empirical facts in the whole field of quantitative economics." The graph below is from Gibson's 1926 piece.

However, the correlation was first noted by Thomas Tooke (for more see Pivetti's entry in the New Palgrave; subscription required), the leader of the Banking School, in the 19th century, and author of the massive and underappreciated History of Prices ( Vols. 1, 2, 3, 4, 5 and 6 available on-line).

In that book, and contrary to Ricardo and the Bullionist authors (the precursors of the Currency School), he argued that inflation during the Napoleonic Wars (1793-1814) was not caused by money printing by the Bank of England (bank notes were inconvertible from 1797 to 1821), but the result of bad crops and higher prices of agricultural goods, import restrictions associated to the blockade and higher import prices, and exchange rate depreciation that added to the costs of imported goods. The last cause he cites is associated to what Keynes called the Gibson Paradox (p. 347):

"A higher rate of interest, in consequence of the absorption by the war loans of a considerable proportion of the savings of individuals; such higher rate of interest constituting an increased cost of production."

In other words, the rate of interest enters the costs of production [some heterodox authors, like Lance Taylor, have referred to this as the Cavallo-Patman effect]. Note that there is no particular paradox in the effect, which according to Lawrence Klein [yes the same that won the Sveriges Riksbank Prize, aka the Nobel] is still doing fine (and yes also needs subscription; sorry).

The reason that the correlation seemed like a paradox to Keynes is because he thought along marginalist lines and assumed that a higher rate of interest would lead to lower investment, and hence reduce demand pressures on prices. He expected a negative correlation between the two variables. Wicksell noted that if the economy was hit by a real shock (to the marginal productivity of capital, for example) then you would have a bank rate of interest lower than the new and higher natural rate of interest (implying higher remuneration for the more productive capital), and the excess investment associated to this situation would lead to higher prices. Eventually, banks would note that the bank rate was too low, and would raise it, leading to an increase in the rate of interest together with higher prices (a positive correlation). Of course Wicksell and the natural rate cannot survive the capital critique. For more on Wicksell go here.