Hírek

How to Forecast the Business Cycle? Sentiment Speaks

As a central question for academics, policy analysis, and financial practice, the causality between Main Street and Wall Street is controversial and still fiercely debated. This blog article does not intend to elaborate more on the relationship between the financial and business cycles. Instead, assuming the former (potentially) causes the latter, policymakers are more interested in extrapolating and chewing over the predictive ability of the asset markets on future macroeconomic performance and exploring the underlying transmission channels.

López-Salido et al. (2017) (hereinafter LZS) attempt to shed some empirical light on the above question. They examine the forecasting power of the credit market sentiment, measured by the change of the credit spread (ΔCS) between the seasonally adjusted long-term Baa-rated industrial bonds and the treasury bond yield, on the future GDP growth. Figure 1 shows the credit spread over the period from 1925 and 2013, presenting an explicit countercyclical nature as the credit spread widens ahead and during each recession (represented by the shaded area, defined by NBER). Running a preliminary one-step predictive regression on future GDP using the current ΔCS as the predictor, LZS find the predictive ability of the change of credit spread is substantial. A 90 basis points increase in credit spread declines GDP growth by 1.8 percentage points for the next period.

Baa-Treasury Credit Spread

Data source: López-Salido et al. (2017)

Can we be sure that it is the causal effect of investor sentiment on the real economy that explains this result? Presumably, we can decompose ΔCS into two components. One represents past investor sentiment, and another contains the expectation of changes in future cashflows. Note that only the first component can be interpreted as a cause of future fluctuations in real activity, as the second one is driven by simply the anticipation of them. With the above simple one-step prediction, it is impossible to determine which component dominates the comovement. To this end, LZS propose using a two-step regression to isolate the first component from the other and bypass any identification issue. In other words, the proposed two-step design can be used to test the predictive power of the component of ΔCS that is only driven by past investor sentiment.

To be more specific, in the first stage, LZS use variables reflecting investor sentiment, namely, the valuation indicator, the high-yield bond issuance and the Baa-Treasury credit spread level, to forecast the following period credit spread change. Then, in the second stage, they employ the fitted ΔCS to predict future output growth. The results show that a move from the 25th to the 75th quantile of the fitted ΔCS reduces the real GDP growth in the next period by two percentage points, suggesting the prior sentiment is the primary explainer. More interestingly, in the first stage, the sign of the coefficient for the credit spread level is negative. This implies a mean-reverse dynamic in the credit spread. Motivated by the empirical findings from the two-step estimation, LZS propose their main conjecture:  The mean reversal of the credit spread at period t-1 pulls back the booming sentiment in the previous period and thus lowers the credit supply, causing a contraction in actual output in period t.

However, without further discussion, above credit-market supply-side hypothesis can be easily overturned by an alternative story highlighting that the sentiment proxies may predict something about the future demand rather than the supply. To disentangle and examine their hypothesis, LZS build a model to flesh out the implication of the credit supply channel, by employing the change in the composition of external finance as a critical identification between the supply- and demand-side story. Intuitively, an increase in the credit supply should surge the net debt issuance relative to the net equity issuance, which, in turn, is not influenced by the credit demand change. Fitting the model with actual data, LZS find empirical evidence in favor of the credit supply story.

Taken together, the elevated sentiment in the credit markets today widens the credit spread tomorrow. It demotivates the credit supply from lower credit-quality firms, exerting a future contraction in the output. Through this mechanism, the past credit sentiment paves a new but crucial road to map future economic activity. This channel also offers some insights into monetary policy. Noticeably, any policy-induced change in the credit-risk premium today will ultimately lead to an opposite direction in two years. If so, the central bank needs to consider the nontrivial intertemporal tradeoff and be more careful in implementing aggressive policies.

Chaoyi Chen


References:

López-Salido, D., Stein, J. C., & Zakrajšek, E. (2017). Credit-market sentiment and the business cycle. The Quarterly Journal of Economics, 132(3), 1373–1426.


Főoldali kép forrása: pixabay.com

The post How to Forecast the Business Cycle? Sentiment Speaks appeared first on Economania blog.

Ez a weboldal sütiket („cookie”) használ
Ez a weboldal sütiket használ a kényelmesebb böngészés érdekében. A honlap használatával Ön elfogadja, hogy az oldal sütiket használ. Kérjük, olvassa el Sütitájékoztatónkat, amelyben további információkat olvashat a sütikről és azt is megtudhatja, hogyan tudja kikapcsolni vagy törölni őket. View more
Cookies settings
Elfogadom
Nem fogadom el
Adatvédelmi és Cookie szabályzat
Privacy & Cookies policy
Cookie name Active
Hogyan törölheti a cookie-kat, és hogyan tilthatja le azokat: Kérjük, olvassa el Sütitájékoztatónkat, amelyben további információkat olvashat a sütikről és azt is megtudhatja, hogyan tudja kikapcsolni vagy törölni őket. Amennyiben nem szeretné, hogy cookie-kat használjunk, letilthatja azokat. A letiltás böngészőfüggő, és különböző módon történhet. A legnépszerűbb böngészőkben a letiltás mikéntjéről az alábbi linkeken szerezhet tudomást:
Save settings
Cookies settings