Appendix
Data construction and source
Business confidence index We obtain the business confidence index from the OECD’s leading indicator database. The OECD collects business confidence data, based on business tendency survey of manufacturing activity, from the Institute for Supply Management (ISM).Footnote 21 The business confidence series refers to PMI (previously, PMI referred to the Purchasing Managers’ Index), which is based on Manufacturing ROB. The PMI is an equally weighted (20% each) composite index of five seasonally adjusted diffusion indices, namely new orders, production, employment, supplier deliveries and inventories. An index value of over 50 represents growth or expansion within the manufacturing sector of the economy compared with the prior month and a value of under 50 indicates contraction. The OECD converts the PMI diffusion index into a net balance (in %) for cross-country consistency.
Real business investment and its components The real business investment corresponds to the private non-residential fixed investment and its components are non-residential structure, equipment and intellectual property products. We obtain the data from NIPA Table 1.1.3 of BEA. Real gross domestic product We obtain the data for the real gross domestic product from NIPA Table 1.1.3 of BEA.
Price index of gross domestic product We obtain the data for the price index of real gross domestic product from NIPA Table 1.1.4 of BEA.
Price index of business investment We obtain the data for the price index of real business investment from NIPA Table 1.1.4 of BEA.
Real lending rate It is the prime business rate of commercial bank. We obtain the data from Economic Research Division, Federal Reserve Bank of St. Louis. Source: Board of Governors of the Federal Reserve System.
We calculate the real lending rate as an ex post measure as follows:
$$\begin{aligned} R_t = (i/100) - \log (\text {Price index of}\, \text {GDP}_{t+1}/\text {Price index of}\, \text {GDP}_{t}) \end{aligned}$$
(17)
where R is the real lending rate.
User cost of capital We measure the user cost of capital following Chirinko and Schaller (2001) and Ang (2010), which is similar to the Hall and Jorgenson (1969). The user cost of capital is as follows:
$$\begin{aligned} \text {CC}_t = (R_t + \text {DEP}_t)(\text {Price index of}\, \text {TBI}_t/\text {Price index of}\, \text {GDP}_t) \end{aligned}$$
(18)
where DEP is the depreciation. We fix the DEP as 5%.
Real cash flow It is the net cash flow with Inventory Valuation Adjustment (IVA) divided by the price index of gross domestic product. We obtain from Economic Research Division, Federal Reserve Bank of St. Louis. Source: BEA.
Stock market price It is the monthly S&P 500 index divided by the price index of gross domestic product. We collect from Yahoo!Finance.Footnote 22
Term spread It is the monthly rate of 10-year government bond minus the monthly rate of 3-month treasury bill. We obtain from Economic Research Division, Federal Reserve Bank of St. Louis. Source: Board of Governors of the Federal Reserve System.
Credit spread It is the Moody’s Baa corporate bond yield minus the Moody’s Aaa corporate bond yield. We obtain from Economic Research Division, Federal Reserve Bank of St. Louis. Source: Board of Governors of the Federal Reserve System.
SR calculation
The calculation of SR is as:
$$\begin{aligned} SR = \frac{{\hat{g}}^{uu}+{\hat{g}}^{dd}}{{\hat{g}}^{uu}+{\hat{g}}^{du}+{\hat{g}}^{ud}+{\hat{g}}^{dd}}, \end{aligned}$$
(19)
where \(\hat{g_t}\), u and d are the forecast of \(g_t\), upward signal and downward signal, respectively.
$$\begin{aligned} {\hat{g}}^{uu}= & {} \sum _{t=1}^{T}\mathbf{1 }[\hat{g_t} =1, g_t =1],\\ {\hat{g}}^{ud}= & {} \sum _{t=1}^{T}\mathbf{1 }[\hat{g_t} =1, g_t =0],\\ {\hat{g}}^{du}= & {} \sum _{t=1}^{T}\mathbf{1 }[\hat{g_t} =0, g_t =1],\\ {\hat{g}}^{dd}= & {} \sum _{t=1}^{T}\mathbf{1 }[\hat{g_t} =0, g_t =0] \end{aligned}$$
Table 13 OOS results: predictive ability of BCI for investment downturns, using control variables Additional models for downturns and direction of investment
Table 13 contains the results to assess the robustness whether BCI has independent forecasting power for business investment downturns, after controlling for other relevant predictors. Panel (a) shows that BCI-nested model performs better than BCI non-nested model for 1–4-quarter horizons. The result suggests that BCI has additional information to forecast investment downturns, after controlling for conventional predictors, TS and \(\Delta \hbox {SP}\) of recessions. Panel (b) shows that BCI-nested model is superior than BCI non-nested model for all forecast horizons, where we control for CS and \(\Delta \hbox {SP}\). We next consider CS, \(\Delta \hbox {SP}\) and \(\Delta \hbox {GDP}\) as control variables and show the results in panel (c). This result is also consistent with previous result and implies that BCI has independent information to forecast the investment downturns for 1–4-quarter horizons.
Table 14 OOS results: predictive ability of BCI for direction of investment growth, using control variables Table 15 Baseline forecast of business investment growth Finally, we use different control variables to evaluate whether BCI forecast for direction of investment growth independently. Table 14 shows the results. Panel (a) shows that BCI-nested model is better than BCI non-nested model for 1- and 3-quarter forecast horizons, suggesting that BCI has independent information to explain the direction of investment, controlling for conventional predictors, TS and \(\Delta \hbox {SP}\). We then control for CS and \(\Delta \hbox {SP}\) and show the results in panel (b). The results show that BCI-nested model has better performance than BCI non-nested model for 1- and 4-quarter horizons. Panel (c) also shows the results after controlling for three predictors, CS, \(\Delta \hbox {SP}\) and \(\Delta \hbox {GDP}\) and suggests that BCI has additional information to explain the direction of investment for 2-quarter horizons (Table 15).