Managerial Overconfidence, Conservative Accounting and Corporate Investment Keiichi Shima and Junichi Nakamura Available at SSRN, Id: 3203187

Backgrounds: Investment efficiencies Keiichi Shima (C)

Firm Value Investment Maximum Optimum Under- investment Over- investment Both destroy the firm value Keiichi Shima (C)

Managerial Overconfidence Heaton (2002), Hackbarth (2008, 2009): Overestimate the return / Underestimate the risk, Make upward biased cash flow forecasts, Exercise real options earlier Overinvestment Believe their stock is undervalued in capital market, Avoid external financing / Prefer internal cash (likely) Underinvestment Keiichi Shima (C)

Conservative Accounting Timely recognition of losses and delayed recognition of gains (Basu, 1997) Curb managers’ opportunistic behavior / More timely signaling of default risk to creditors, thereby lower debt costs (Watts, 2003; Zhang, 2008) Recover efficiencies Underinvestment in positive NPVs, due to higher gains verification / Decrease managerial incentives to make riskier investments (Roychowdhury, 2010; Kravet, 2014) Keiichi Shima (C)

Research Design (Baseline: Abel and Eberly, 1994) Keiichi Shima (C)

( , ) represent earnings; Shock, follows a GBM: = + is a standard Wiener process Capital stock, follows: = ( − ) Keiichi Shima (C)

Firm value: ( , ) = max [න 0 ∞ − [(+ , + ) − (+ , + )]] The Bellman equation: (, ) = max [(, ) − (, ) + 1 []] Keiichi Shima (C)

Define the shadow price of capital: ≡ The Bellman equation becomes (, ) = max [(, ) − (, ) + ( − ) + + 22 ] Then, the firm's optimal investment: ∗ ∈ arg max [ − (, )] Keiichi Shima (C)

The adjustment cost function for firm at time (Eberly, 1997): (, , ) = + 1 + ( ) 1+ The acquisition price of capital: The firm's optimal investment function: ln = 1 ln ( + 2 + 3 ) + 4 + Keiichi Shima (C)

The extensions: Cash Flow & Real Options Keiichi Shima (C)

ln = 1 ln + 2 + 3 + 1 + 2 + 3 + 4 + ෍ + ෍ + Keiichi Shima (C)

The extensions: Managerial Overconfidence & Conservative Accounting Keiichi Shima (C)

ln = 1 ln + 2 + 3 +1 + 2 + 3 + 4 + 5 + 6 + 7 + 8 + 9 + 10 + 11 + 12 + 13 + 14 + ෍ + ෍ + Keiichi Shima (C)

Managerial Overconfidence Proxy Keiichi Shima (C)

Lin et al.'s (2005) idea to construct a firm-year managerial overconfidence proxy Classify firms as overconfident if they have more upward- biased forecasts than downward-biased forecasts over past 5 years Employ ex-ante and ex-post measures of upward-bias of earnings forecasts to construct our managerial overconfidence proxy Keiichi Shima (C)

Keiichi Shima (C) t-1 t+1 Actual Earnings(t) t Earnings Forecast(t) > Ex-post upward-bias, − Earnings Forecast(t+1) Ex-ante upward-bias, < ∩

Management earnings forecasts at time as upwardly- biased if −1 = 1, i.e., firms missed their previous forecasts and think that their earnings will increase this period Averaging −1 for the past five years: = ෍ =0 4 − −1− /5 OC dummy if OCindex>.5 Keiichi Shima (C)

Conservative Accounting Proxy Keiichi Shima (C)

Calculate C_Score (Khan and Watts, 2009) i.e., asymmetric timeliness of earnings in reflecting economic losses relative to economic gains _ = መ 1 + መ 2 + መ 3 + መ 4 log of equity market value, market to book value of equity, long- and short-term debt/equity market value Keiichi Shima (C)

Rank _ each year and standardize them ranging between 0 and 1 Denote the annual rank of _ ∈ 0,1 , by Then, calculate the five-year average of : = ෍ =1 5 − /5 CONS dummy if CSindex>median Keiichi Shima (C)

Real Option Variables Proxy Keiichi Shima (C)

Use real sales growth for earnings growth proxy μit is calculated as the average growth rate of real sales over the past five years σit is calculated as the standard deviation of the real sales growth over the past five years Keiichi Shima (C)

Data Sources Keiichi Shima (C)

Public firms’ BS and PL: DBJ Corporate Financial Databank Equity price and management earnings forecasts: NEEDS FinancialQuest Estimation period between April 2000 and March 2015 All variables winsorized at 1 and 99 percent (25,856 firm- year observations) Keiichi Shima (C)

Results Keiichi Shima (C)

ln = 1 ln + 2 + 3 +1 + 2 + 3 + 4 + 5 + 6 + 7 + 8 + 9 + 10 + 11 + 12 + 13 + 14 + ෍ + ෍ + Keiichi Shima (C)

Parameter α1 0.565 (0.331)* α2 -9.841 (3.293)*** α3 7.943 (3.218)** Keiichi Shima (C) 1 ln + 2 + 3

Cash flow sensitivity CFit 0.349 (0.043)*** Negit CFit -0.421 (0.132)*** Real options sensitivity μit 2.162 (0.214)*** σit -0.848 (0.205)*** Keiichi Shima (C) +1 + 2 + 3 + 4

Keiichi Shima (C) CFit OCit 0.175 (0.056)*** Negit CFit OCit -0.366 (0.157)** μit OCit 0.275 (0.258) σit OCit -0.294 (0.252) + 5 + 6 + 7 + 8

Keiichi Shima (C) CFit Consit 0.436 (0.053) Negit CFit Consit -0.223 (0.154) μit Consit 1.350 (0.243)*** σit Consit 0.054 (0.259) + 10 + 11 + 12 + 13

OCit -0.111 (0.305)*** Consit -0.231 (0.031)*** Keiichi Shima (C) OC and Cons Fixed Effects:

5 + 6 + 7 + 8 + 9 Cashflow * Overconfidence: Positive significant Negative significant Real Options * Overconfidence: , Insignificant Fixed Effect: Negative significant OC exhibits greater and riskier cash flow sensitivity / Likely underinvestment fixed effect /No real options related effect Keiichi Shima (C)

10 +11 +12 +13 +14 Cashflow * Conservatism: , Insignificant Real Options * Conservatism: Positive significant Insignificant Fixed Effect: Negative significant Cons exhibits no cash flow effect / Conservative firms with higher growth invest more / Underinvestment fixed effect Keiichi Shima (C)