That current challenge to monetary policy
It is reassuring to read that the Bangko Sentral ng Pilipinas (BSP) continues to sharpen its analytical tools particularly on predicting potential financial crisis and presenting a more detailed inflation outlook for more reliable forecasting. A good handle on the future is the heart and soul of the BSP’s flexible inflation targeting.
Reassuring, yes, because the BSP has not ceased producing more and more analytical tools for the macroeconomy and stress testing methodologies for surveillance of the banking system. Quite a mouthful, but all of these initiatives should further enhance the batting average of the BSP in projecting inflation and deciding on the most fitting monetary policy stance.
To be hawkish, or to be dovish, is not a question of fashion, or cut and paste, but a result of careful assessment of both past, current, and future data. It’s obvious that when the BSP can confidently manage price movements in the Philippines, we benefit from such positive consequences as more sustainable economic growth, more competitive exchange rate, and a more investment-friendly economic environment. Price stability also reinforces the economy’s resiliency to both global and domestic shocks.
But there’s something that BSP Governor Eli Remolona cited recently that should further keep the BSP economists and researchers busy in the next few years. He said: “We need to do what I call narrative identification. It’s not enough to have an index that’s based on averages. We have to somehow calibrate them so that they actually predict the crisis, especially our own crisis.”
The narrative approach, if applied to macroeconomic identification, uses qualitative sources of data and information like, but not limited to, newspapers, government records, or, in the case of central banks like the BSP, transcripts of Monetary Board meetings. This ragtag group of sources can provide information crucial in establishing causal relationships.
The husband-wife team of David H. Romer and Christina D. Romer, both of the University of California, Berkeley, employed this methodology first in 1989 in looking into the impact of monetary policy by reading the historical minutes and transcripts of the Federal Reserve Board meetings on monetary policy and incorporating these to statistical analysis. They opted for this approach to deal with the common problem of omitted variable bias to consider a more complete set of relevant explanatory factors.
In her January 2023 presidential address before the American Economic Association in New Orleans, Ms. Romer explained that their methodology allowed them to identify precisely “a subset of monetary actions that were not motivated by other factors affecting output.” The behavior of output following the US Fed monetary shocks “would provide relatively unbiased estimates of the impact of monetary policy.”
For economists, what the Romers are saying is that there are other relevant variables that do affect both output and the motivation of the US Fed to change the course of monetary policy. One common flaw, and this is something that both economists and non-economists are guilty of, is when one concludes that monetary policy does not matter when we see it morphing from hawk to dove without any perceptible effect on business activities.
Auspicious as it was, Ms. Romer’s presidential address presented her with the opportunity to revisit their 1989 paper in the context of the recent issue of the US Fed monetary policy allegedly motivating a possible recession in the US. After over three decades, the husband-wife team has learned volumes “about the pitfalls of narrative research” and how to turn it around.
Ms. Romer cited the key ingredients to a more meaningful use of the narrative approach in order to avoid being just “literary,” a term used to discredit their initial 1989 paper. First is the reliability of their source. Their 1989 paper was based on the abridged record of policy actions of the Federal Open Market Committee, or to some, minutes of those Committee meetings. This means they extracted very little information on the motivations of the Fed’s policy actions. With a shorter lag involved before these minutes were made public, the Romers felt that they were “less forthright.”
Their new research is anchored on “very detailed summaries of the discussion with extensive paraphrases, or verbatim transcripts.” As Ms. Romer stressed, they are contemporaneous with the Fed’s monetary policy decisions. With a long lag before the transcripts’ publication, the US Fed governors must have been more open and more forthright.
The second key to a good narrative approach is a clear sense of what one wants to extract from the source. Citing the genius of the iconic Milton Friedman and Anna Schwartz in employing the same methodology in their seminal work, A Monetary History of the United States (1963), proponents of the methodology should focus on those instances when monetary policy was undertaken without regard to economic activities, and the effects are expected to be unbiased estimates of causality between monetary policy and economic performance.
One drawback to this general approach is coming up with specific criteria to be used in judging for the period to be considered. It is difficult to tell whether the economy is at, or close to, potential output but it is perhaps even more difficult to determine whether the corresponding inflation target is just right, or needed some adjustment. Ms. Romer then argued that there is a scope for including not only contractionary monetary policy shocks but also expansionary monetary policy shocks. If monetary policy matters, output should be affected upwards.
The team sought to identify significant contractionary and expansionary changes in monetary policy that were not exactly considered to address real sector activities in the United States from 1946 to 2016. Based on this approach, they succeeded in identifying 10 instances when the Fed deliberately adjusted monetary policy to alter the path of business activities. There was only one case where the Fed adjusted monetary policy in response to high inflation. This is not surprising because, as Ms. Romer pointed out, no monetary shock was observed between 1988 and 2016. Even as they included almost 30 years of additional observations, inflation in the US had been steady during that period, until recently.
The Romers’ revisit confirmed most of their original findings with a few tweaks here and there on periodization and interpretations. Building on Friedman and Schwartz’ pioneering work, the Romers incorporated their narrative evidence into a more rigorous statistical technique, the Jordà local projection approach (2005). They admit that there are other statistical techniques available to achieve rigor and unbiased estimates. It was also shown that alternative specifications yielded only minor differences in the results. The results tested robustly.
How does the Romers latest study interpret current monetary conditions and US Fed policy?
For one thing, even as the transcripts of the US Fed monetary policy meetings will not be available until 2028, a reading of the narrative evidence based on the record of such meetings would indicate that the US Fed tightened monetary policy in response to recordhigh inflation in the last couple of years. The US Fed, on the record, announced that this was unacceptable, so monetary policy must respond. Fed funds target rate has risen a full four percentage points, or 400 basis points, something akin to the 1988 monetary policy action by the US central bank.
Ms. Romer, in her presidential address, was quite fearless in her prognosis of what to expect from such monetary policy action. She said we should not expect inflation to fall rapidly. It’s possible perhaps after two quarters after the shock. If the supply side is favorable, a more rapid decline is likely. From January 2023, the impact on unemployment should be felt throughout the same year. Finally, Ms. Romer suggested that the US Fed was facing a difficult decision on when to stop tightening monetary policy and start reducing the policy rate.
With long and variable lags of monetary actions, it is possible the US Fed might still be tightening until they see a more definitive trend decline in inflation. But that would signal that “they have gone farther than they needed to.” Ms. Romer admitted it’s an impossible call to say how much more monetary policy has to continue, and how much longer interest rates should remain high.
In our previous columns, we made the point that the BSP was correct in fighting inflation until it returns to the 2-4% target. Some of the expected effects of this monetary policy shock include economic moderation and weaker job creation. There is no way by which those factors affecting inflation could not affect output and the labor market.
Yet, we have seen that as the BSP maintains higher interest rate longer, inflation rates in the Philippines seem to have started to gravitate towards the 2-4% inflation target recently. Yet, it must be recognized that the impact on both output and jobs has not been as bad as some quarters expect. The Philippines is expecting output growth close to the lower end of the 6-7% target while the latest jobs statistics indicate lower unemployment and underemployment.
It would bring more completeness to the analysis of inflation and inflation prospects if the BSP would start implementing the call of the Governor in doing the narrative approach in addressing the usual identification problem in macroeconomic research. While other alternative techniques are available, the narrative approach is very promising, though difficult. For instance, it would require the Monetary Board greater understanding for more transparency and forthrightness in allowing the publication of the transcripts of monetary policy meetings. A longer lag may be considered to permit more open and frank discussion of monetary policy.
But we are optimistic that once this is done, inflation management in the Philippines would have a very interesting narrative.