Essays on US business cycle investigated from Keynesian perspective

Kim, Minsu (2023) Essays on US business cycle investigated from Keynesian perspective. PhD thesis, University of Nottingham.

PDF (Thesis - as examined) - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Available under Licence Creative Commons Attribution.
Download (7MB) | Preview


This thesis investigates the US business cycle from the (old) Keynesian perspective. It is in line with the [86]Keynes (1936) in following aspects. (1) The analyses are focused on the cyclical fluctuations. The underlying assumption behind this practice is that economic fluctuations are mainly ascribed to cyclical components. (2) Business cycles are only analysed using demand-side factors. This practice corresponds to the Keynes’ proposition that demand generates supply. (3) The main mechanism of business cycles is based on the [86]Keynes’(1936) intuition that investment is driven by businessmen’s expectations and the major source of the expectations is current consumption.

Three chapters are connected to each other. In chapter 3, I detrend the US real GDP to obtain a reasonable business cycle. In this course, I detect three break points in the trend. In chapter 4, I investigate the main driver of business cycles using the data which is detrended using the break points detected in chapter 3. In chapter 5, I apply the mechanism that is studied in chapter 4 to forecasting.

In chapter 2, I review the literature of modern business cycle models, that is, dynamic stochastic general equilibrium models (DSGE) from an empirical perspective which is corresponding to the Old Keynesian idea. I discuss the limitation of DSGE models for the two empirical facts, consumption-investment comovement and flat Phillips curve.

In chapter 3, I decompose the US real GDP using a deterministic (log) linear trend with three breaks. The practice is chosen as it yields a cycle that is important in magnitude, which corresponds to the Keynesian perspective in that economic fluctuations are mainly ascribed to a cyclical component. The measured business cycle displays a boom-bust pattern with sporadic downward pluckings. The estimated boom-bust cycle is far from a stationary process in that it is long-lasting, large in size and displays sharp reversions to the trend. To support the estimated result, I provide following evidence: (1) I also estimate the cyclical components of well-known US coincident indicators using the detected break points. The estimated cycles of US coincident indicators are close to the estimated cycle of real GDP. (2) the US business cycles estimated using the HP filter with large smoothing parameter and the bandpass filter with an alternative bandwidth also corresponds to the cycle estimated through a linear deterministic trend with breaks. (3) When applying the same estimation strategy, boom-bust cycles are found in Korea and Japan real GDP. To justify the estimated results, I also show that the estimated trend and cycle provide a new and reasonable view for the well-known macroeconomic puzzle: the slow recovery of the US economy from the 2008-9 Global Financial Crisis. The estimation results of this chapter imply that the Global Financial Crisis is the collapse of a huge boom rather than a big recession, therefore, a recovery does not follow. Meanwhile, non-stationary cycles provide an important implication in the business cycle literature. They allow for a long-lasting and large business cycle observed in reality without unrealistically large shocks or endogenous amplification mechanisms.

In chapter 4, I investigate the main driver of business cycles using the data which is detrended using the break points detected in chapter 1. I argue that consumption shocks are the main driver of business cycles from the perspective of [86]Keynes (1936) and [115]Pigou (1927). I reconcile their ideas based on their commonality highlighting the role of businessmen’s expectations as an important driver of business cycles. The reconciled idea describes a boom-bust cycle generated by consumption shocks and their propagation to investment. To examine the idea, I estimate the effects of consumption shocks. The estimated effects predict a a boombust cycle close to the benchmark investment cycle. A unit shock to consumption generates a long-lasting and important responses of consumption and investment even though the secular trend of each variable is already removed. This result is corresponding to boom-bust cycles but in contrast with a well-known humpshaped responses of output to a demand shock. The studied model is close to the Keynes-Hansen-Samuelson multiplier accelerator model ([122]Samuelson, 1939) in that investment is the function of consumption changes. The main differences are in two points: (1) marginal effect of output changes on consumption is smaller than in the Keynesian cross. (2) Consumption shocks are the most important among demand shocks.

In chapter 5, I apply the mechanism that is studied in chapter 4 to forecasting. I forecast short-term US GDP growth. The trend growth is forecasted using the HP filter with a large smoothing parameter. The cyclical changes (=demeaned output growth) are forecasted using the consumption-investment relationship that is studied in chapter 4. The forecasting performance of the model is comparable to the one of the Survey of Professional Forecasters.

Item Type: Thesis (University of Nottingham only) (PhD)
Supervisors: Zu, Yang
Omar, Licandro
Keywords: Keynesian economics, US business cycle, dynamic stochastic general equilibrium models, DSGE, real GDP
Subjects: H Social sciences > HB Economic theory
Faculties/Schools: UK Campuses > Faculty of Social Sciences, Law and Education > School of Economics
Item ID: 72021
Depositing User: KIM, Minsu
Date Deposited: 04 Jan 2024 11:12
Last Modified: 04 Jan 2024 11:12

Actions (Archive Staff Only)

Edit View Edit View