The Cost Channel of Monetary Transmission

This 2001 NBER Macroeconomics Annual paper by Marvin Barth and Valerie Ramey tests whether monetary policy operates through a supply-side “cost channel” – in which a monetary contraction raises firms’ working-capital financing costs and thereby raises production costs and prices – in addition to the conventional demand channel, using two separate empirical frameworks. The first is an aggregate six-variable quarterly VAR (1959:Q1-2000:Q3) that uses long-run restrictions following Gali (1999) to identify a technology shock (the only shock permitted a permanent effect on labor productivity) and a monetary policy shock (the federal-funds-rate innovation orthogonal to it); the authors show that the estimated responses to a contractionary monetary shock – falling labor productivity, falling hours, procyclically falling real wages, and a temporary price-level increase – resemble the responses to a negative technology shock far more than the responses to an identified demand shock (a defense-spending increase), which instead raises output and lowers real wages countercyclically. The second framework is a seven-equation industry-level pseudo-panel VAR (monthly, 1959-2000, 7 lags, with a macro block of industrial production, the PCE deflator, commodity prices, M2, and the federal funds rate, and an industry block of output and the price-to-wage ratio for 21 two- and three-digit manufacturing industries) that tests whether a contractionary funds-rate shock raises the industry price-to-wage ratio (P/W) – the signature of a supply-side cost shock – even as output falls. Over the full sample, 10 of 21 industries show P/W rising for more than half of the first 24 months, 6 significantly at the 10% level; the effect is far stronger before 1979 (15 of 21 industries significant) than during 1983-2000 (only 3 of 21 significant), a period split the authors attribute to financial deregulation, the end of Federal Reserve “credit actions” that had directly restricted bank lending, and the shift to floating exchange rates. Consistent with this timing, in a standardized 25-basis-point funds-rate shock the pre-Volcker output trough is almost four times as deep and the “price puzzle” (prices rising after a contractionary shock) persists for over two years, versus a roughly nine-month funds-rate reversion and largely flat prices after 1983. The authors motivate the mechanism with a working-capital calculation (Federal Reserve Flow of Funds and BEA data, averaged 1959-2000): gross working capital equals about 17 months of final sales and net working capital about 11 months, and a simple labor-demand equation implies that, holding real wages fixed, a 100-basis-point rate rise lowers labor demand by about 3% when the capital cost share is 0.3 – or by about 15% once a credit-channel-style widening of external financing spreads is layered onto the roughly 400-basis-point average funds-rate increase seen over a Romer tightening episode. Cross-industry heterogeneity in the price-to-wage response correlates positively with industries’ interest-expense burden relative to sales (correlation of about 0.53 in an early-1970s cross-section, about 0.43 in a 1990 cross-section), which the authors read as further evidence linking the price response to financing costs specifically. They test and argue against sticky-price/flexible-wage and misspecified-Fed-reaction-function explanations for the same facts, but explicitly acknowledge that countercyclical-markup theories are observationally equivalent to their cost-channel story without direct marginal-cost data, and they frame the cost channel as a short-run phenomenon layered on top of the standard demand channel rather than a challenge to long-run monetary neutrality.

Marvin J. Barth III | Macro Paper Warehouse

The Cost Channel of Monetary Transmission