predictors.csv

predictor,code_file,Acronym,Cat.Signal,Predictability in OP,Signal Rep Quality,Authors,Year,LongDescription,Journal,Cat.Form,Cat.Data,Cat.Economic,SampleStartYear,SampleEndYear,Acronym2,Sign,Stock Weight,LS Quantile,Quantile Filter,Filter,Detailed Definition
abnormal_accruals,abnormal_accruals.py,AbnormalAccruals,Predictor,1_clear,2_fair,Xie,2001,Abnormal Accruals,AR,continuous,Accounting,accruals,1971,1992,AccrAbn,-1,EW,0.1,NA,NA,"Define Accruals as net income (ib) minus operating cash flow (oancf), divided by average total assets (at) for years t-1 and t. If oancf is missing, replace operating cash flow with funds from operations (fopt) minus the annual change in total current assets (act) plus the annual change in cash and short-term investments (che) plus the annual change in current liabilities (lct) minus the annual change in debt in current liabilities (dlc). For each year t and 2-digit sic code, regress Accruals on: the inverse of average total assets for year t-1, the change in revenue (sale) from year t-1 to t divided by total assets for t-1, propery plant and equipment (ppegt) divided by total assets for t-1.  AbnormalAccrual is the residual from this cross-sectional regression.  See code for more details."
abnormal_accruals_pct,abnormal_accruals.py,,Predictor,1_clear,2_fair,Xie,2001,Abnormal Accruals,AR,continuous,Accounting,accruals,1971,1992,,-1,EW,0.1,NA,NA,"Define Accruals as net income (ib) minus operating cash flow (oancf), divided by average total assets (at) for years t-1 and t. If oancf is missing, replace operating cash flow with funds from operations (fopt) minus the annual change in total current assets (act) plus the annual change in cash and short-term investments (che) plus the annual change in current liabilities (lct) minus the annual change in debt in current liabilities (dlc). For each year t and 2-digit sic code, regress Accruals on: the inverse of average total assets for year t-1, the change in revenue (sale) from year t-1 to t divided by total assets for t-1, propery plant and equipment (ppegt) divided by total assets for t-1.  AbnormalAccrual is the residual from this cross-sectional regression.  See code for more details."
accruals,accruals.py,Accruals,Predictor,1_clear,1_good,Sloan,1996,Accruals,AR,continuous,Accounting,accruals,1962,1991,Accruals,-1,EW,0.1,NA,abs(prc)>5,Annual change in current total assets (act) minus annual change in cash and short-term investements (che) minus annual change in current liabilities (lct) minus annual change in debt in current liabilities (dlc) minus change in income taxes (txp). All divided by average total assets (at) over this year and last year. Exclude if abs(prc) < 5.
am,am.py,AM,Predictor,1_clear,1_good,Fama and French,1992,Total assets to market,JF,continuous,Accounting,valuation,1963,1990,AM,1,EW,NA,NA,NA,Total assets (at) divided by market value of equity.
ann_ret,ann_ret.py,AnnouncementReturn,Predictor,1_clear,1_good,"Chan, Jegadeesh and Lakonishok",1996,Earnings announcement return,JF,continuous,Price,earnings event,1977,1992,AnnounRet,1,EW,NA,NA,NA,Get announcement date for quarterly earnings from IBES (fpi = 6). AnnouncementReturn is the sum of (ret - mktrf + rf) from one day before an earnings announcement to 2 days after the announcement.
asset_growth,asset_growth.py,AssetGrowth,Predictor,1_clear,1_good,"Cooper, Gulen and Schill",2008,Asset growth,JF,continuous,Accounting,investment,1968,2003,InvestAG,-1,EW,0.1,NA,NA,Annual growth rate of total assets (at)
beta_liq_ps,beta_liq_ps.py,BetaLiquidityPS,Predictor,1_clear,1_good,Pastor and Stambaugh,2003,Pastor-Stambaugh liquidity beta,JPE,continuous,Price,liquidity,1968,1999,BetaLiquidityPS,1,VW,0.1,NA,abs(prc)>5,"Monthly excess return (ret -rf) regressed on innovations in liquidity from Pastor's website (\url{https://faculty.chicagobooth.edu/lubos.pastor/research/liq_data_1962_2018.txt}). Use 60 month rolling window regression, and require at least 36 non-missing observations."
beta_tail_risk,beta_tail_risk.py,BetaTailRisk,Predictor,1_clear,1_good,Kelly and Jiang,2014,Tail risk beta,RFS,continuous,Price,risk,1963,2010,BetaTailRisk,1,EW,0.2,NA,abs(prc)>5,"Each month, compute the 5th percentile over daily returns over all firms. For all daily return observations with return below that 5th percentile, compute the average of (log(ret/5th percentile of cross-sectional return distribution). Call that average tailEX. BetaTailRisk is the coefficient of a 120-month rolling regression of a firm's stock return on tailEX. Exclude if price less than 5 or share code greater than 11."
beta_vix,beta_vix.py,betaVIX,Predictor,1_clear,1_good,Ang et al.,2006,Systematic volatility,JF,continuous,Price,volatility,1986,2000,betaVIX,-1,VW,0.2,NA,NA,Coefficient on daily change in the VIX of a 1-month rolling window regression of daily stock excess returns on market return and the daily change in the CBOE S&P 100 volatility index (downloaded from FRED). Require at least 15 non-missing observations.
bm,bm.py,BM,Predictor,1_clear,1_good,Stattman,1980,"Book to market, original (Stattman 1980)",Other,continuous,Accounting,valuation,1962,1976,BM,1,EW,0.1,NA,"exchcd%in%c(1,2)",Log of tangible book equity (ceqt) over market equity matched at FYE
bm_rrl,bm_rrl.py,,Predictor,1_clear,2_fair,"Rosenberg, Reid, and Lanstein",1985,Book to market using most recent ME,JPM,continuous,Accounting,valuation,1973,1984,,1,EW,0.2,NA,NA,Log of annual book equity (ceq) over market equity.
bm_dec,bm_dec.py,BMdec,Predictor,1_clear,1_good,Fama and French,1992,Book to market using December ME,JPM,continuous,Accounting,valuation,1963,1990,BMdec,1,EW,NA,NA,NA,BM using most recent December value of market equity.
book_leverage,book_leverage.py,BookLeverage,Predictor,1_clear,1_good,Fama and French,1992,Book leverage (annual),JF,continuous,Accounting,leverage,1963,1990,BookLev,-1,EW,NA,NA,NA,"Total assets (at) divided by book value of equity plus deferred taxes (txditc) and preferred stock. Equity is shareholder equity (seq) if available, or book equity (ceq) plus preferred stock (pstk, if missing pstkrv, if missing pstkl), or total assets minus total liabilities (lt)."
bpebm,ebm_bpebm.py,BPEBM,Predictor,1_clear,1_good,"Penman, Richardson and Tuna",2007,Leverage component of BM,JAR,continuous,Accounting,leverage,1963,2001,BMlev,-1,EW,NA,NA,abs(prc)>5,"BP - EBM, where BP = (ceq + tstkp - dvpa)/(shrout*abs(prc)), and EBM is defined above. Exclude if price less than 5."
cash,cash.py,Cash,Predictor,1_clear,1_good,Palazzo,2012,Cash to assets,JFE,continuous,Accounting,asset composition,1972,2009,Cash,1,EW,0.1,NA,NA,Ratio of quarterly cash and short-term investments (cheq) and total assets (atq).
cash_prod,cash_prod.py,CashProd,Predictor,1_clear,1_good,Chandrashekar and Rao,2009,Cash Productivity,WP,continuous,Accounting,profitability alt,1963,2003,CashProd,-1,EW,NA,NA,NA,Calculate market value of equity (mve_c) as absolute price (prc) times number of shares outstanding (shrout). Cash productivity is equal to the difference between mve_c and total assets (at) divided by cash and short-term investments (che).
cb_oper_prof,cb_oper_prof.py,CBOperProf,Predictor,1_clear,1_good,Ball et al.,2016,Cash-based operating profitability,JFE,continuous,Accounting,profitability,1963,2014,ProfCash,1,VW,0.1,NYSE,NA,"Revenue (revt) minus cost (cogs) - (administrative expenses (xsga) - R&D expenses (xrd))  minus annual change in receivables (rect), annual change in investment (invt) and annual change in prepaid expenses, plus annual change in current deferred revenue (drc), long-term deferred revenue (drlt), accounts payable (ap) and accrued expenses (xacc), all divided by total assets (at) in year t. Replace all variables in the numerator with 0 if they are missing. Exclude if share code is greater 11, market value of equity, BM or total assets are missing, or if SIC code between 6000 and 6999."
cf,cf.py,CF,Predictor,1_clear,1_good,"Lakonishok, Shleifer, Vishny",1994,Cash flow to market,JF,continuous,Accounting,valuation,1968,1990,CF2Price,1,EW,0.1,NA,"exchcd%in%c(1,2)",Net income (ib) plus depreciation (dp) divided by market equity. Exclude NASDAQ stocks.
cfp,cfp.py,cfp,Predictor,1_clear,1_good,"Desai, Rajgopal, Venkatachalam",2004,Operating Cash flows to price,AR,continuous,Accounting,valuation,1973,1997,CFOper2Price,1,EW,0.2,NA,NA,"Operating cash-flow (oancf) divided by market value of equity. If operating cash-flow is missing, replace by difference betwee net income (ib) and level of accruals, where the latter is the annual change in current assets (act) minus the annual change in cash and short-term investments (che), minus the annual change in current liabilities (lct) plus the annual change in debt in current liabilities (dlc) plus the annual change in payable income taxes (txp) plus depreciation (dp)."
ch_asset_turnover,ch_asset_turnover.py,ChAssetTurnover,Predictor,1_clear,1_good,Soliman,2008,Change in Asset Turnover,AR,continuous,Accounting,sales growth,1984,2002,ATurnGr,1,EW,NA,NA,abs(prc)>5,Annual change in AssetTurnover (defined above). Exclude if price less than 5.
ch_eq,ch_eq.py,ChEQ,Predictor,1_clear,1_good,Lockwood and Prombutr,2010,Growth in book equity,JFR,continuous,Accounting,investment,1964,2007,BEgrowth,-1,EW,0.2,NA,NA,Ratio of book equity (ceq) to book equity in the previous year. Include only if book equity is positive this year and last year.
ch_inv,ch_inv.py,ChInv,Predictor,1_clear,1_good,Thomas and Zhang,2002,Inventory Growth,RAS,continuous,Accounting,investment alt,1970,1997,Invntory,-1,EW,0.1,NA,NA,12 month change in inventory (invt) divided by average total assets.
ch_inv_ia,ch_inv_ia.py,ChInvIA,Predictor,1_clear,1_good,Abarbanell and Bushee,1998,Change in capital inv (ind adj),AR,continuous,Accounting,investment growth,1974,1988,InvestGr,-1,EW,NA,NA,NA,"Growth in capital expenditure (capx) minus average growth in capital expenditure in the same industry (two-digit SIC). If capx is missing, capital expenditure is defined as the annual change in property, plant and equipment (ppent). Capital expenditure growth is defined as the percentage growth of capx today relative to the average capx over the previous two years (.5*(capx$_{t-1}$ + capx$_{t-2}$), or as percentage growth relative to the previous year only if t-2 is missing."
ch_nncoa,ch_nncoa.py,ChNNCOA,Predictor,1_clear,1_good,Soliman,2008,Change in Net Noncurrent Op Assets,AR,continuous,Accounting,investment alt,1984,2002,ChNNCOA,-1,EW,NA,NA,NA,Twelve-month change in noncurrent operating assets. Noncurrent operating assets is ( (at - act - ivao) - (lt - dlc - dltt) )/at.
ch_nwc,ch_nwc.py,ChNWC,Predictor,1_clear,1_good,Soliman,2008,Change in Net Working Capital,AR,continuous,Accounting,investment alt,1984,2002,NWCgr,-1,EW,NA,NA,abs(prc)>5,Twelve-month change in net working capital. Net working capital is ( (act - che) - (lct - dlc) )/at
ch_tax,ch_tax.py,ChTax,Predictor,1_clear,1_good,Thomas and Zhang,2011,Change in Taxes,JAR,continuous,Accounting,other,1977,2006,TaxGr,1,EW,0.1,NA,NA,"4-quarter change in quarterly total taxes (txtq), scaled by lagged total assets (at)."
comp_equ_iss,comp_equ_iss.py,CompEquIss,Predictor,1_clear,1_good,Daniel and Titman,2006,Composite equity issuance,JF,continuous,Accounting,external financing,1968,2003,CompEquIss,-1,EW,NA,NA,abs(prc)>5,5 year growth rate of market value of equity minus 5 year stock return.
comp_debt_iss,comp_debt_iss.py,CompositeDebtIssuance,Predictor,1_clear,1_good,"Lyandres, Sun and Zhang",2008,Composite debt issuance,RFS,continuous,Accounting,external financing,1970,2005,DebtFinC,-1,EW,0.333333333,NA,NA,Log of long-term debt (dltt) plus debt in current liabilties (dlc) minus log of the same variable 5 years ago.
coskew_acx,coskew_acx.py,CoskewACX,Predictor,1_clear,1_good,"Ang, Chen and Xing",2006,Coskewness using daily returns,RFS,continuous,Price,risk,1963,2001,CoskewACX,-1,EW,0.2,NA,"shrcd<=11, exchcd==1","Signal is the sample counterpart of $E[\tilde{r}_{it} \tilde{r}_{mt}^2]/( SD[\tilde{r}_{it} ] SD[\tilde{r}_{mt}]^2$ where $\tilde{r}_{it}$ is the de-meaned stock return and $\tilde{r}_{mt}$ is the de-meaned market excess return.  Signal is computed using the past year of daily data, and using the NYSE CRSP VW index for the market (dsia), with returns continuously compounded.  See code for details."
del_coa,del_coa.py,DelCOA,Predictor,1_clear,1_good,Richardson et al.,2005,Change in current operating assets,JAE,continuous,Accounting,investment alt,1962,2001,AssetCGr,-1,EW,NA,NA,NA,"Difference in current operating assets (total current assets (act) minus cash and short-term investments (che)) between years t-1 and t, scaled by average total assets (at) in years t-1 and t."
del_col,del_col.py,DelCOL,Predictor,1_clear,1_good,Richardson et al.,2005,Change in current operating liabilities,JAE,continuous,Accounting,external financing,1962,2001,LiabCGr,-1,EW,NA,NA,NA,"Difference in current operating liabilities (total current liabilities (lct) minus debt in current liabilities (dlc)) between years t-1 and t, scaled by average total assets (at) in years t-1 and t."
del_equ,del_equ.py,DelEqu,Predictor,1_clear,1_good,Richardson et al.,2005,Change in equity to assets,JAE,continuous,Accounting,investment,1963,2001,Eq2AGr,-1,EW,NA,NA,NA,"Difference in book equity (ceq) between years t-1 and t, scaled by average total assets (at) in years t-1 and t."
del_finl,del_finl.py,DelFINL,Predictor,1_clear,1_good,Richardson et al.,2005,Change in financial liabilities,JAE,continuous,Accounting,external financing,1962,2001,FinLiabGr,-1,EW,NA,NA,NA,"Difference in financial liabilities (sum of long-term debt (dltt), current liabilitites (dlc) and preferred stock (pstk)) between years t-1 and t, scaled by average total assets (at) in years t-1 and t."
del_lti,del_lti.py,DelLTI,Predictor,1_clear,1_good,Richardson et al.,2005,Change in long-term investment,JAE,continuous,Accounting,investment,1962,2001,ChLTI,-1,EW,NA,NA,NA,"Difference in investment and advances (ivao) between years t-1 and t, scaled by average total assets (at) in years t-1 and t."
del_net_fin,del_net_fin.py,DelNetFin,Predictor,1_clear,1_good,Richardson et al.,2005,Change in net financial assets,JAE,continuous,Accounting,investment alt,1962,2001,DelNetFin,1,EW,NA,NA,NA,"Compute the sum of short-term investments (ivst) and investments and advances (ivao) minus the sum of long-term debt (dltt), debt in current liabilities (dlc) and preferred stock capital. Divide the difference between the current and one-year lagged sum by total assets (at) averaged over the current and previous fiscal years."
dnoa,dnoa.py,dNoa,Predictor,1_clear,1_good,"Hirshleifer, Hou, Teoh, Zhang",2004,change in net operating assets,JAE,continuous,Accounting,investment,1964,2002,dNoa,-1,EW,NA,NA,NA,"12-month growth in Net Operating Assets scaled by lagged total assets (at). Net Operating assets are operating assets minus operating liabilities. Operating assets are total assets (at) minus cash- and short-term investments (che), operating liabilities are total assets minus long-term debt (dltt), minority interest (mib), deferred charges (dlc), book equity (ceq) and preferred stock (pstk), all items (except at and ceq) replaced with 0 if missing."
dol_vol,dol_vol.py,DolVol,Predictor,1_clear,1_good,"Brennan, Chordia, Subra",1998,Past trading volume,JFE,continuous,Trading,volume,1966,1995,VolumeDol,-1,EW,NA,NA,NA,Log of two-month lagged trading volume (vol) times two-month lagged price (prc).
earnings_consis,earnings_consis.py,EarningsConsistency,Predictor,1_clear,2_fair,Alwathainani,2009,Earnings consistency,BAR,continuous,Accounting,earnings growth,1971,2002,EarnCons,1,EW,NA,NA,abs(prc)>1,"Average earnings growth over previous 48 months. Earnings growth is defined as EPS (epspx) minus EPS 12 months ago divided by average EPS 12 and 24 months ago. Exclude if price less than 5, absolute value of 12 month earnings growth greater 600%, or earnings growth and earnings growth 12 months ago have different signs."
earnings_surp,earnings_surp.py,EarningsSurprise,Predictor,1_clear,2_fair,"Foster, Olsen and Shevlin",1984,Earnings Surprise,AR,continuous,Accounting,earnings growth,1974,1981,EarnSurp,1,EW,0.1,NA,abs(prc)>5,"EPS (epspxq) minus EPS twelve months ago - Drift, scaled by standard deviation of that expression. Drift is the average earnings growth (EPS - EPS twelve months ago) over the past two years. Exclude if price less than 5"
ebm,ebm_bpebm.py,EBM,Predictor,1_clear,1_good,"Penman, Richardson and Tuna",2007,Enterprise component of BM,JAR,continuous,Accounting,valuation,1963,2001,BMent,1,EW,0.1,NA,abs(prc)>5,(ceq + che - dltt - dlc - dc - dvpa+ tstkp ) / (mve\_c + che - dltt - dlc - dc - dvpa+ tstkp). Exclude if price less than 5.
ent_mult,ent_mult.py,EntMult,Predictor,1_clear,1_good,Loughran and Wellman,2011,Enterprise Multiple,JFQA,continuous,Accounting,valuation,1963,2009,EntMult,-1,EW,0.1,NA,NA,"Market value of equity + long-term debt (dltt) + debt in current liabilities (dlc) + deferred charges (dc) - cash and short-term investments (che) , divided by operating income (oibdp). Exclude if missing book equity or negative operating income."
ep,ep.py,EP,Predictor,1_clear,1_good,Basu,1977,Earnings-to-Price Ratio,JF,continuous,Accounting,valuation,1957,1971,EP,1,EW,0.2,NA,exchcd==1,"ib / lag(market value of equity, 6 months).  NYSE stocks only.  Exclude if EP < 0.  Lag simulates the Dec 31 market equity used in original paper"
equity_duration,equity_duration.py,EquityDuration,Predictor,1_clear,1_good,"Dechow, Sloan and Soliman",2004,Equity Duration,RAS,continuous,Accounting,valuation,1962,1998,Duration,-1,VW,0.2,NA,NA,see code
firm_age_mom,firm_age_mom.py,FirmAgeMom,Predictor,1_clear,1_good,Zhang,2006,Firm Age - Momentum,JF,continuous,Price,momentum,1983,2001,MomYoung,1,EW,0.2,NA,NA,"6 month return, restricted to the bottom quintile of the cross-sectional firm age distribution. Exclude if price less than 5 or firm younger than 12 months."
frontier,frontier.py,Frontier,Predictor,1_clear,1_good,Nguyen and Swanson,2009,Efficient frontier index,JFQA,continuous,Accounting,valuation,1980,2003,EffFrontier,1,EW,0.1,NA,"exchcd%in%c(1,2,3),shrcd<=11","Frontier is the residual of a regression of log(BM) on log(book equity (ceq)), long-term debt (dltt) to assets (at), capital expenditures (capx) to revenue (sale), R&D expense (xrd) to revenue, advertising expense (xad) to revenue, property plant and equipment (ppent) to assets, EBIT (ebitda) to assets, and dummies for Fama-French's 48 industry definitions. Regression is updated each month with a rolling window of 60 months."
gp,gp.py,GP,Predictor,1_clear,1_good,Novy-Marx,2013,gross profits / total assets,JFE,continuous,Accounting,profitability,1963,2010,ProfGross,1,VW,0.2,NA,NA,"Revenue (sale) - cost of goods solds (cogs), divided by  total assets (at).  Drop if financial."
gr_ad_exp,gr_ad_exp.py,GrAdExp,Predictor,1_clear,1_good,Lou,2014,Growth in advertising expenses,RFS,continuous,Accounting,investment alt,1974,2010,AdExpGr,-1,EW,0.1,NA,abs(prc)>5,"Log of advertising expense (xad) minus log of advertising expense last year. Exclude if price less than 5, xad less than .1 or stock in the lowest decile of market value of equity."
gr_capx,gr_capx.py,grcapx,Predictor,1_clear,1_good,Anderson and Garcia-Feijoo,2006,Change in capex (two years),JF,continuous,Accounting,investment growth,1976,1999,CAPXgr,-1,EW,0.2,NA,NA,"Growth rate of capital expenditures (capx) relative to two years ago.   If capx is missing, replace with annual change in property, plant and equipment (ppent)."
gr_capx_1yr,gr_capx.py,,Predictor,1_clear,1_good,Anderson and Garcia-Feijoo,2006,Change in capex (two years),JF,continuous,Accounting,investment growth,1976,1999,,-1,EW,0.2,NA,NA,"Growth rate of capital expenditures (capx) relative to two years ago.   If capx is missing, replace with annual change in property, plant and equipment (ppent)."
gr_capx_3yr,gr_capx.py,grcapx3y,Predictor,1_clear,1_good,Anderson and Garcia-Feijoo,2006,Change in capex (three years),JF,continuous,Accounting,investment growth,1976,1999,CAPXgr3y,-1,EW,0.2,NA,NA,"Capital expenditures (capx) divided by the sum of capital expenditures from year - 1, year -2, and year -3.   If capx is missing, replace with annual change in property, plant and equipment (ppent)."
herf,herf.py,Herf,Predictor,1_clear,1_good,Hou and Robinson,2006,Industry concentration (sales),JF,continuous,Other,other,1963,2001,Herf,-1,EW,0.2,NA,NA,"Three-year rolling average of the three digit industry Herfindahl index based on firm revenue (sale). Exclude regulated industries (4011, 4210, 4213 & year $\leq$ 1980; 4512 & year $\leq$ 1978, 4812, 4813 & year $\leq$ 1982, 4900-4999 in any year)"
herf_be,herf_be.py,HerfBE,Predictor,1_clear,1_good,Hou and Robinson,2006,Industry concentration (equity),JF,continuous,Other,other,1963,2001,HerfBE,-1,EW,0.2,NA,NA,"Three-year rolling average of the three digit industry Herfindahl index based on firm book equity. Exclude regulated industries (4011, 4210, 4213 & year $\leq$ 1980; 4512 & year $\leq$ 1978, 4812, 4813 & year $\leq$ 1982, 4900-4999 in any year)"
hire,hire.py,hire,Predictor,1_clear,1_good,"Bazdresch, Belo and Lin",2014,Employment growth,JPE,continuous,Other,investment alt,1965,2010,LaborGr,-1,EW,0.2,NA,NA,"Change in number of employees (emp) between t -1 and t, scaled by average number of employees in t-1 and t. Replace hire with 0 if emp or lagged emp is missing."
idio_vol_ff3,idio_vol_ff3.py,IdioVol3F,Predictor,1_clear,1_good,Ang et al.,2006,Idiosyncratic risk (3 factor),JF,continuous,Price,volatility,1963,2000,IdioVol3F,-1,VW,0.2,NA,NA,Standard deviation of residuals from Fama-French three factor regressions using the past month of daily data.  Value weighted
idio_vol_aht,idio_vol_aht.py,IdioVolAHT,Predictor,1_clear,1_good,"Ali, Hwang, and Trombley",2003,Idiosyncratic risk (AHT),JFE,continuous,Price,volatility,1976,1997,IdioVolAHT,-1,VW,NA,NA,NA,Standard deviation of residuals from CAPM regressions using the past year of daily data. Require at least 100 non-missing observations.
illiquidity,illiquidity.py,Illiquidity,Predictor,1_clear,1_good,Amihud,2002,Amihud's illiquidity,JFM,continuous,Trading,liquidity,1964,1997,Illiquid,1,EW,NA,NA,"abs(prc)>5, exchcd==1",Past twelve month average of: daily return (abs(ret)) divided by turnover((abs(prc)*vol)
intan_bm,intan.py,IntanBM,Predictor,1_clear,1_good,Daniel and Titman,2006,Intangible return using BM,JF,continuous,Accounting,long term reversal,1968,2003,IntanBM,-1,EW,NA,NA,abs(prc)>5,"In each month, run a cross-sectional regression of a firm's five-year stock return on 5 year lagged BM (defined above) and a constructed regressor that is the change in BM from 5 years ago to today plus the five-year stock return. The residual from that regression is IntanBM."
intan_cfp,intan.py,IntanCFP,Predictor,1_clear,1_good,Daniel and Titman,2006,Intangible return using CFtoP,JF,continuous,Accounting,long term reversal,1968,2003,IntanCFP,-1,EW,NA,NA,abs(prc)>5,"In each month, run a cross-sectional regression of a firm's five-year stock return on the 5 year lagged  CFP = (net income (ni) plus depreciation (dp))/market value of equity and a constructed regressor that is the change in CFP from 5 years ago to today plus the five-year stock return. The residual from that regression is IntanCFP."
intan_ep,intan.py,IntanEP,Predictor,1_clear,1_good,Daniel and Titman,2006,Intangible return using EP,JF,continuous,Accounting,long term reversal,1968,2003,IntanEP,-1,EW,NA,NA,abs(prc)>5,"In each month, run a cross-sectional regression of a firm's five-year stock return on the 5 year lagged  EP = net income (ni)/market value of equity and a constructed regressor that is the change in EP from 5 years ago to today plus the five-year stock return. The residual from that regression is IntanEP."
intan_sp,intan.py,IntanSP,Predictor,1_clear,2_fair,Daniel and Titman,2006,Intangible return using Sale2P,JF,continuous,Accounting,long term reversal,1968,2003,IntanSP,-1,EW,NA,NA,NA,"In each month, run a cross-sectional regression of a firm's five-year stock return on 5 year lagged SP (defined above) and a constructed regressor that is the change in SP from 5 years ago to today plus the five-year stock return. The residual from that regression is IntanSP."
int_mom,int_mom.py,IntMom,Predictor,1_clear,1_good,Novy-Marx,2012,Intermediate Momentum,JFE,continuous,Price,momentum,1927,2010,Mom12to7,1,VW,0.1,NA,NA,Stock return between months t-12 and t-6
investment,investment.py,Investment,Predictor,1_clear,1_good,"Titman, Wei and Xie",2004,Investment to revenue,JFQA,continuous,Accounting,investment,1973,1996,InvToRev,-1,VW,0.2,NA,NA,Ratio of capital investment (capx) to revenue (revt) divided by the firm-specific 36-month rolling mean of that ratio. Exclude if revenue less than \$10m.
invest_ppe_inv,invest_ppe_inv.py,InvestPPEInv,Predictor,1_clear,1_good,"Lyandres, Sun and Zhang",2008,change in ppe and inv/assets,RFS,continuous,Accounting,investment,1970,2005,InvestPPEInv,-1,EW,0.3,NA,NA,"One-year change in property, plants and equipment (ppegt) plus one year change in inventory (invt), scaled by one-year lagged assets (at)."
inv_growth,inv_growth.py,InvGrowth,Predictor,1_clear,1_good,Belo and Lin,2012,Inventory Growth,RFS,continuous,Accounting,profitability,1965,2009,InvenGr,-1,EW,0.1,NA,NA,"Defate invt growth using gnp deflator.  Signal is deflated invt growth rate from fiscal year t to fiscal year t-1.  Drop if 1 digit sic code is 4 or 6, or if at or ppent <= 0"
leverage,leverage.py,Leverage,Predictor,1_clear,1_good,Bhandari,1988,Market leverage,JF,continuous,Accounting,leverage,1952,1981,Leverage,1,EW,NA,NA,NA,Total liabilities (lt) divided by market value of equity.
lt_reversal,lt_reversal.py,LRreversal,Predictor,1_clear,1_good,De Bondt and Thaler,1985,Long-run reversal,JF,continuous,Price,long term reversal,1929,1982,Mom36m,-1,EW,NA,NA,NA,Stock return between months t-36 and t-13.
max_ret,max_ret.py,MaxRet,Predictor,1_clear,1_good,"Bali, Cakici, and Whitelaw",2011,Maximum return over month,JFE,continuous,Price,volatility,1962,2005,MaxRet,-1,VW,0.1,NA,NA,Maximum of daily returns (ret) over the previous month
mean_rank_rev_growth,mean_rank_rev_growth.py,MeanRankRevGrowth,Predictor,1_clear,2_fair,"Lakonishok, Shleifer, Vishny",1994,Revenue Growth Rank,JF,continuous,Accounting,sales growth,1968,1990,RevGrowth,1,EW,0.2,NA,"exchcd%in%c(1,2)","Rank firms by their annual revenue growth each year over the past 5 years. MeanRankRevGrowth is the weighted average of ranks over the past 5 years, that is, MeanRankRevGrowth = (5*Rank$_{t-1}$ + 4*Rank$_{t-2}$ + 3*Rank$_{t-3}$ + 2*Rank$_{t-4}$ + 1*Rank$_{t-5}$)/15. Exclude NASDAQ stocks."
mom12m,mom12m.py,Mom12m,Predictor,1_clear,1_good,Jegadeesh and Titman,1993,Momentum (12 month),JF,continuous,Price,momentum,1964,1989,Mom12m,1,EW,0.1,NA,NA,Stock return between months t-12 and t-1
mom12m_off_season,mom12m_off_season.py,Mom12mOffSeason,Predictor,1_clear,1_good,Heston and Sadka,2008,Momentum without the seasonal part,JFE,continuous,Price,other,1965,2002,Mom12mOffSeason,1,EW,0.1,NA,"exchcd%in%c(1,2)",Average return in  other months over the previous year.
mom6m,mom6m.py,Mom6m,Predictor,1_clear,1_good,Jegadeesh and Titman,1993,Momentum (6 month),JF,continuous,Price,momentum,1964,1989,Mom6m,1,EW,0.1,NA,NA,Stock return between months t-6 and t-1
mom6m_junk,mom6m_junk.py,Mom6mJunk,Predictor,1_clear,1_good,Avramov et al,2007,Junk Stock Momentum,JF,continuous,Price,momentum,1985,2003,Mom6Jnk,1,EW,0.1,NA,"abs(prc)>5, me>me_nyse20",Mom6m. Include only stocks with a credit rating (splticrm) of BBB or lower
mom_off_season,mom_off_season.py,MomOffSeason,Predictor,1_clear,1_good,Heston and Sadka,2008,Off season long-term reversal,JFE,continuous,Price,other,1965,2002,MomOffSeason,-1,EW,0.1,NA,"exchcd%in%c(1,2)",Average return in other months over the preceding 2-5 years.
mom_off_season_6yr_plus,mom_off_season_6yr_plus.py,MomOffSeason06YrPlus,Predictor,1_clear,1_good,Heston and Sadka,2008,Off season reversal years 6 to 10,JFE,continuous,Price,other,1965,2002,MomOffSeason6YrPlus,-1,EW,0.1,NA,"exchcd%in%c(1,2)",Average return in other months over the preceding 6-10 years.
mom_off_season_16yr_plus,mom_off_season_16yr_plus.py,MomOffSeason16YrPlus,Predictor,1_clear,1_good,Heston and Sadka,2008,Off season reversal years 16 to 20,JFE,continuous,Price,other,1965,2002,MomOffSeason16YrPlus,-1,EW,0.1,NA,"exchcd%in%c(1,2)",Average return in other months over the preceding 16-20 years.
mom_season,mom_season.py,MomSeason,Predictor,1_clear,1_good,Heston and Sadka,2008,Return seasonality years 2 to 5,JFE,continuous,Price,other,1965,2002,MomSeason,1,EW,0.1,NA,"exchcd%in%c(1,2)",Average return in the same month over the preceding 2-5 years.
mom_season_6yr_plus,mom_season_6yr_plus.py,MomSeason06YrPlus,Predictor,1_clear,1_good,Heston and Sadka,2008,Return seasonality years 6 to 10,JFE,continuous,Price,other,1965,2002,MomSeason06YrPlus,1,EW,0.1,NA,"exchcd%in%c(1,2)",Average return in the same month over the preceding 6-10 years.
mom_season_11yr_plus,mom_season_11yr_plus.py,MomSeason11YrPlus,Predictor,1_clear,1_good,Heston and Sadka,2008,Return seasonality years 11 to 15,JFE,continuous,Price,other,1965,2002,MomSeason11YrPlus,1,EW,0.1,NA,"exchcd%in%c(1,2)",Average return in the same month over the preceding 11-15 years.
mom_season_16yr_plus,mom_season_16yr_plus.py,MomSeason16YrPlus,Predictor,1_clear,1_good,Heston and Sadka,2008,Return seasonality years 16 to 20,JFE,continuous,Price,other,1965,2002,MomSeason16YrPlus,1,EW,0.1,NA,"exchcd%in%c(1,2)",Average return in the same month over the preceding 16-20 years.
mom_season_short,mom_season_short.py,MomSeasonShort,Predictor,1_clear,1_good,Heston and Sadka,2008,Return seasonality last year,JFE,continuous,Price,other,1965,2002,MomSeasonShort,1,EW,0.1,NA,"exchcd%in%c(1,2)",Average return in the same monthin the previous year.
net_debt_fin,net_debt_fin.py,NetDebtFinance,Predictor,1_clear,1_good,"Bradshaw, Richardson, Sloan",2006,Net debt financing,JAE,continuous,Accounting,external financing,1971,2000,NDebtFin,-1,EW,0.1,NA,NA,"Long-term debt issuance (dltis) minus long-term debt reduction (dltr) minus current debt changes (dlcch), scaled by average total assets (at) in years t-1 and t. Replace missing values of dlcch with 0. Exclude if ratio is greater than 1."
net_debt_price,net_debt_price.py,NetDebtPrice,Predictor,1_clear,1_good,"Penman, Richardson and Tuna",2007,Net debt to price,JAR,continuous,Accounting,leverage,1963,2001,NDebtPrice,-1,EW,0.2,NA,NA,"Long-term debt (dltt) plus debt in current liabilities (dlc) plus preferred stock (pstk) plus preferred dividends in arrears (dvpa) minus treasury stock (tstkp) minus cash and short-term investments (che), scaled by market value of equity. Exclude if SIC between 6000 and 6999, or if missing value for total assets (at), net income (ib), common shares outstanding (csho), book value of equity (ceq) or price close fiscal year (prcc\_f).  Keep only 3rd B/M Quintile, following Table 4 (and in contrast to Table 1)."
net_equity_fin,net_equity_fin.py,NetEquityFinance,Predictor,1_clear,1_good,"Bradshaw, Richardson, Sloan",2006,Net equity financing,JAE,continuous,Accounting,external financing,1971,2000,NEqFin,-1,EW,0.1,NA,NA,"Sale of common stock (sstk) minus  purchase of common stock (prstkc), scaled by average total assets (at) from years t and t-1. Exclude if absolute value of ratio is greater than 1."
net_payout_yield,net_payout_yield.py,NetPayoutYield,Predictor,1_clear,2_fair,Boudoukh et al.,2007,Net Payout Yield,JF,continuous,Accounting,valuation,1984,2003,NPayYield,1,EW,0.1,NA,NA,"Dividends (dvc) plus purchase of common and preferred stock (prstkc) minus sale of common and preferred stock (sstk), divided by market value of equity lagged 6 months.  Exclude if NetPayoutYield is 0, financial firm based on SIC code, ceq <= 0, or less than 2 years in CRSP"
noa,noa.py,NOA,Predictor,1_clear,1_good,Hirshleifer et al.,2004,Net Operating Assets,JAE,continuous,Accounting,asset composition,1964,2002,NOA,-1,EW,0.1,NA,NA,"Difference between operating assets and operating liabilities, scaled by lagged total assets. Operating assets are total assets (at) minus cash- and short-term investments (che), operating liabilities are total assets minus long-term debt (dltt), minority interest (mib), deferred charges (dc) and book equity (ceq)."
op_leverage,op_leverage.py,OPLeverage,Predictor,1_clear,1_good,Novy-Marx,2011,Operating leverage,ROF,continuous,Accounting,other,1963,2008,OperLeverage,1,EW,0.2,NA,NA,"Sum of administrative expenses (xsga) and cost of goods sold (cogs), scaled by total assets (at).  Use xsga = 0 if xsga is missing."
order_backlog,order_backlog.py,OrderBacklog,Predictor,1_clear,1_good,"Rajgopal, Shevlin, Venkatachalam",2003,Order backlog,RAS,continuous,Accounting,sales growth,1981,1999,OrderBacklog,-1,EW,NA,NA,NA,Order backlog (ob) divided by average total assets (at) in years t-1 and t. Exclude if order backlog is 0.
order_backlog_chg,order_backlog_chg.py,OrderBacklogChg,Predictor,1_clear,2_fair,Baik and Ahn,2007,Change in order backlog,Other,continuous,Accounting,accruals,1971,1999,OrderBacklogChg,1,EW,0.1,NA,NA,Define normalized order backlog  as order backlog (ob) divided by average total assets (at) in years t-1 and t. Exclude if order backlog is 0.  Signal is normalized order backlog minus normalized order backlog one year ago.
org_cap,org_cap.py,OrgCap,Predictor,1_clear,1_good,Eisfeldt and Papanikolaou,2013,Organizational capital,JF,continuous,Accounting,R&D,1970,2008,OrgCap,1,VW,0.2,NA,NA,"Defined recursively. Set xsga to 0 if missing.  Initialize with OrgCap = 4*general expenses (xsga) in the first year, and calculate as .85*OrgCap previous year + xsga/gnpdeflator current year thereafter. Scale by total assets (at).  Keep only december year ends, and non-financials, and non-missing sic code.  Then winsoriz at the 1% level and standardized by subtracting the same month within-industry mean (Fama French 17 industries) and dividing by the same month within industry standard deviation of OrgCap."
org_cap_noadj,org_cap.py,,Predictor,1_clear,1_good,Eisfeldt and Papanikolaou,2013,Organizational capital,JF,continuous,Accounting,R&D,1970,2008,,1,VW,0.2,NA,NA,"Defined recursively. Set xsga to 0 if missing.  Initialize with OrgCap = 4*general expenses (xsga) in the first year, and calculate as .85*OrgCap previous year + xsga/gnpdeflator current year thereafter. Scale by total assets (at).  Keep only december year ends, and non-financials, and non-missing sic code.  Then winsoriz at the 1% level and standardized by subtracting the same month within-industry mean (Fama French 17 industries) and dividing by the same month within industry standard deviation of OrgCap."
payout_yield,payout_yield.py,PayoutYield,Predictor,1_clear,2_fair,Boudoukh et al.,2007,Payout Yield,JF,continuous,Accounting,valuation,1984,2003,PayYield,1,EW,0.1,NA,NA,"Sum of dividends (dvc), purchase of common and preferred stock (prstkc) and max(preferred stock redemption value (pstkrv), 0) over market value of equity lagged 6 months.  Exclude if PayoutYield $\leq$ 0, financial firm based on SIC code, ceq <= 0, or less than 2 years in CRSP"
pct_acc,pct_acc.py,PctAcc,Predictor,1_clear,1_good,"Hafzalla, Lundholm, Van Winkle",2011,Percent Operating Accruals,AR,continuous,Accounting,accruals,1989,2008,AccrOper,-1,EW,0.1,NA,abs(prc)>5,"Income before extraordinary items (ib) minus net cash flow (oancf) divided by absolute value of ib. If oancf is missing, PctAcc is defined as ( (act - act$_{t-12}$) - (che - che$_{t-12}$) - (  (lct - lct$_{t-12}$) - (dlc - dlc$_{t-12}$) - (txp - txp$_{t-12}$) - dp ) )/abs(ib). In either case, if ib is equal to 0, divide by .01 instead. Exclude if price less than 5."
pct_tot_acc,pct_tot_acc.py,PctTotAcc,Predictor,1_clear,1_good,"Hafzalla, Lundholm, Van Winkle",2011,Percent Total Accruals,AR,continuous,Accounting,accruals,1989,2008,AccrPct,-1,EW,0.1,NA,NA,"Net income (ni) minus (purchase of common and preferred stock (prstkcc) minus sale of common and preferred stock (sstk) plus dividends (dvt), cash flow from operations (oancf), from financing (fincf) and investment (ivncf)). Scaled by absolute value of net income."
price,price.py,Price,Predictor,1_clear,1_good,Blume and Husic,1973,Price,JF,continuous,Price,other,1932,1971,Price,-1,EW,NA,NA,NA,Log of absolute value of price (prc).
rd,rd.py,RD,Predictor,1_clear,1_good,"Chan, Lakonishok and Sougiannis",2001,R&D over market cap,JF,continuous,Accounting,R&D,1975,1995,RD,1,EW,0.2,NA,NA,R&D expense (xrd) over market value of equity.
rd_ability,rd_ability.py,RDAbility,Predictor,1_clear,1_good,"Cohen, Diether and Malloy",2013,R&D ability,RFS,continuous,Accounting,other,1980,2009,RDAbility,1,EW,0.2,NA,abs(prc)>5,"Regress log of sales growth (sale over sale in previous year) on log of (1+ xrd/sale) in 5 bivariate regressions with (1+xrd/sale) lagged by $1, \ldots, 5$ years. Run regressions over previous 8 years and require at least 6 non-missing observations. Also require at least half of past research and development observations to be non-zero. RDAbility is the mean of the coefficients on the five lags of log(1+xrd/sale). Set to missing if firm is not in the highest tercile of xrd/sale in a year or if reseach and development expenses are non-positive."
rd_cap,rd_cap.py,RDcap,Predictor,1_clear,1_good,Li,2011,R&D capital-to-assets,RFS,continuous,Accounting,asset composition,1980,2007,RDcap,1,VW,0.5,NA,NA,R&D capital to assets is the weighted sum of lagged R&D expenditures scaled by the assets (at). We replace xrd with 0 if xrd is missing and compute the numerator as $xrd_t + .8 xrd_{t-1}+ .6xrd_{t-2} + .4 xrd_{t-3} + .2 xrd_{t-4}$. Replace RDcap with missing before  1980 or if firm is in upper two thirds of market cap (shrout*abs(prc)) distribution in a month.
realized_vol,idio_vol_ff3.py,RealizedVol,Predictor,1_clear,1_good,Ang et al.,2006,Realized (Total) Volatility,JF,continuous,Price,volatility,1963,2000,IdioVol,-1,VW,0.2,NA,NA,Standard deviation of residuals from CAPM regressions using the past month of daily data.  Value weighted
resid_mom6m_ff3,resid_mom.py,,Predictor,1_clear,1_good,"Blitz, Huij and Martens",2011,Momentum based on FF3 residuals,JEmpFin,continuous,Price,momentum,1930,2009,,1,EW,0.1,NA,abs(prc)>1,"Run a rolling regression over 36 months of excess return (retrf) on excess market return (mktrf), size and value factors (smb, hml) and compute idiosyncratic returns as the one-month lagged residual. ResidualMomentum is the rolling mean of the residual divided by the rolling standard deviation of the residual, both computed over the past 11 months."
resid_mom12m_ff3,resid_mom.py,ResidualMomentum,Predictor,1_clear,1_good,"Blitz, Huij and Martens",2011,Momentum based on FF3 residuals,JEmpFin,continuous,Price,momentum,1930,2009,MomResid,1,EW,0.1,NA,abs(prc)>1,"Run a rolling regression over 36 months of excess return (retrf) on excess market return (mktrf), size and value factors (smb, hml) and compute idiosyncratic returns as the one-month lagged residual. ResidualMomentum is the rolling mean of the residual divided by the rolling standard deviation of the residual, both computed over the past 11 months."
return_skew,return_skew.py,ReturnSkew,Predictor,1_clear,1_good,"Bali, Engle and Murray",2015,Return skewness,Book,continuous,Price,risk,1963,2012,RetSkew,-1,EW,0.2,NA,NA,Skewness of daily returns (ret) over previous month.
return_skew_ff3,idio_vol_ff3.py,ReturnSkew3F,Predictor,1_clear,1_good,"Bali, Engle and Murray",2015,Idiosyncratic skewness (3F model),Book,continuous,Price,risk,1963,2012,RetSkew3F,-1,EW,0.2,NA,NA,"Skewness of idiosyncratic returns computed as residuals from regression of daily excess returns (ret - rf) on Fama-French factors (mktrf, smb, hml) over the previous month. We require at least 15 non-missing observations."
revenue_surp,revenue_surp.py,RevenueSurprise,Predictor,1_clear,2_fair,Jegadeesh and Livnat,2006,Revenue Surprise,JAE,continuous,Accounting,sales growth,1987,2003,RevSurprise,1,EW,0.2,NA,NA,Define revenue per share as quarterly revenue (revtq) divided by quarterly common shares outstanding (cshprq). RevenueSurprise is the 4-quarter change in revenue per share minus the average 4-quarter change in revenue per share over the previous 2 years. RevenueSurprise is scaled by its standard deviation over the previous 2 years. Exclude if price less than 5.
roa_q,roa_q,roaq,Predictor,1_clear,1_good,"Balakrishnan, Bartov and Faurel",2010,Return on assets (qtrly),JAE,continuous,Accounting,profitability,1976,2005,RoAq,1,EW,0.1,NA,abs(prc)>1,Quarterly net income (ibq) divided by lagged total assets (atq).
share_iss_1yr,share_iss_1yr.py,ShareIss1Y,Predictor,1_clear,1_good,Pontiff and Woodgate,2008,Share issuance (1 year),JF,continuous,Accounting,external financing,1970,2003,ShareIs5,-1,EW,NA,NA,NA,Growth in number of shares between t-18 and t-6. Number of shares is calculated as shrout/cfacshr to adjust for splits.
share_iss_5yr,share_iss_5yr.py,ShareIss5Y,Predictor,1_clear,2_fair,Daniel and Titman,2006,Share issuance (5 year),JF,continuous,Accounting,external financing,1968,2003,ShareIs1,-1,EW,NA,NA,NA,5-year growth in number of shares. Number of shares is calculated as shrout/cfacshr to adjust for splits.
size,size.py,Size,Predictor,1_clear,1_good,Banz,1981,Size,JFE,continuous,Price,size,1926,1975,Size,-1,EW,0.5,NA,NA,Log of monthly market value of equity (abs(prc)*shrout)).
std_turn,std_turn.py,std_turn,Predictor,1_clear,1_good,"Chordia, Subra, Anshuman",2001,Share turnover volatility,JFE,continuous,Trading,liquidity,1966,1995,TurnovVol,-1,EW,NA,NA,exchcd==1,Standard deviation of turnover (vol/shrout) over the past 36 months.
st_reversal,st_reversal.py,STreversal,Predictor,1_clear,1_good,Jegadeesh,1990,Short term reversal,JF,continuous,Price,short-term reversal,1934,1987,Mom1m,-1,EW,0.1,NA,NA,Stock return (ret) over the previous month.
tax,tax.py,Tax,Predictor,1_clear,1_good,Lev and Nissim,2004,Taxable income to income,AR,continuous,Accounting,other,1973,2000,Tax2E,1,EW,NA,NA,abs(prc)>5,"Ratio of Taxes paid and tax share of net income. Numerator is defined as the sum of foreign (txfo) and federal (txfed) income taxes. If either one is missing, numerator is defined as total taxes (txt) minus deferred taxes (txdi). Denominator is the product of the prevailing tax rate and net income (ib). Tax rate is .48 before 1979, .46 from 1979 to 1986, .4 in 1987, .34 between 1988 and 1992 and .35 from 1993 onwards. If net income is negative, and the numerator is positive, tax is defined as 1. Exclude if price less than 5."
total_accruals,total_accruals.py,TotalAccruals,Predictor,1_clear,1_good,Richardson et al.,2005,Total accruals,JAE,continuous,Accounting,investment alt,1962,2001,TotalAccruals,-1,EW,NA,NA,NA,"Before 1988: Change in net working capital ((act - che) - (lct - dlc)) plus change in net noncurrent assets ( (at - act - ivao) - (lt - dlc - dltt)) plus change in net financial assets ( (ivst + ivao - (dltt + dlc + pstk)). Starting in 1988: net income (ni) minus total, operating and investment cashflows (oancf, ivncf, fincf) plus stock sales minus repurchases and dividends (sstk, prstkc, dv)). Scaled by lagged total assets (at). Replace missings in ivao, ivst, dltt, dlc, pstk sstk, prstkc, dv with 0."
trend_factor,trend_factor.py,TrendFactor,Predictor,1_clear,1_good,"Han, Zhou, Zhu",2016,Trend Factor,JFE,continuous,Price,momentum,1930,2014,TrendFactor,1,EW,0.2,NA,NA,See paper section 2.1 and 2.2
vol_std,vol_std.py,VolSD,Predictor,1_clear,1_good,"Chordia, Subra, Anshuman",2001,Volume Variance,JFE,continuous,Trading,liquidity,1966,1995,VolumeSD,-1,EW,NA,NA,exchcd==1,Rolling standard deviation of monthly trading volume (vol) over the past 36 months (require at least 24 observations). Include only NYSE stocks.
x_fin,x_fin.py,XFIN,Predictor,1_clear,1_good,"Bradshaw, Richardson, Sloan",2006,Net external financing,JAE,continuous,Accounting,external financing,1971,2000,ExtFinNet,-1,EW,0.1,NA,NA,Sale of common stock (sstk) minus dividends (dv) minus purchase of common stock (prstkc) plus long-term debt issuance (dltis) minus long-term debt reductions (dltr). Scaled by total assets (at).
zero_trade6m,zero_trade.py,zerotrade,Predictor,1_clear,1_good,Liu,2006,Days with zero trades,JFE,continuous,Trading,liquidity,1960,2003,ZeroTrade,1,EW,0.1,NA,exchcd==1,"In each month, count the number of days with no trades. Define zerotrade as the number of days without trades plus (the sum of monthly turnover (vol/shrout) divided by 48*10$^5$), multiplied by 21/number of trading days per month. Zerotrade is the 6-month average of that variable."
zero_trade1m,zero_trade.py,zerotradeAlt1,Predictor,1_clear,1_good,Liu,2006,Days with zero trades,JFE,continuous,Trading,liquidity,1960,2003,zerotradeAlt1,1,EW,0.1,NA,exchcd==1,"In each month, count the number of days with no trades. Define zerotrade as the number of days without trades plus (the sum of monthly turnover (vol/shrout) divided by 48*10$^5$), multiplied by 21/number of trading days per month. Take 1-month average."
zero_trade12m,zero_trade.py,zerotradeAlt12,Predictor,1_clear,1_good,Liu,2006,Days with zero trades,JFE,continuous,Trading,liquidity,1960,2003,zerotradeAlt12,1,EW,0.2,NA,exchcd==1,"In each month, count the number of days with no trades. Define zerotrade as the number of days without trades plus (the sum of monthly turnover (vol/shrout) divided by 48*10$^5$), multiplied by 21/number of trading days per month. Take 12-month average."
ad_exp,ad_exp.py,AdExp,Predictor,2_likely,1_good,"Chan, Lakonishok and Sougiannis",2001,Advertising Expense,JF,continuous,Accounting,R&D,1975,1996,AdExp,1,EW,0.2,NA,NA,Advertising expense (xad) over market value of equity (shrout*abs(prc))
beta,beta.py,Beta,Predictor,2_likely,1_good,Fama and MacBeth,1973,CAPM beta,JPE,continuous,Price,risk,1929,1968,Beta,1,EW,NA,NA,NA,Coefficient of a 60-month rolling window regression of monthly stock returns minus the riskfree rate on market return minus the risk free rate (ewretd - rf). Exclude if estimate based on less than 20 months of returns.
beta_fp,beta_fp.py,BetaFP,Predictor,2_likely,3_distant,Frazzini and Pedersen,2014,Frazzini-Pedersen Beta,JFE,continuous,Price,other,1929,2012,BetaFP,1,EW,0.1,NA,NA,"Using daily data, call tempRi the sum of today's return and the past 2 days, call tempRm the sum of today's mkt return and the past 2 days mkt return.  Regress return on tempRi and tempRm using past 5 years of data, with a minimum of 3 years.  BetaFP is the square root of the r square from this regression times the ratio of the idiosyncratic stock vol and the market vol."
coskew,coskew.py,Coskewness,Predictor,2_likely,1_good,Harvey and Siddique,2000,Coskewness,JF,continuous,Price,risk,1964,1993,Coskew,-1,VW,0.3,NA,shrcd<=11,"Signal is the sample counterpart of $E[\tilde{r}_{it} \tilde{r}_{mt}^2]/( SD[\tilde{r}_{it} ] SD[\tilde{r}_{mt}]^2$ where $\tilde{r}_{it}$ is the de-meaned stock excess return and $\tilde{r}_{mt}$ is the de-meaned market excess return.  Signal is computed using the past 60 months of monthly data, and using the NYSE/AMEX CRSP VW index for the market (msic).  See code for details."
del_drc,del_drc.py,DelDRC,Predictor,2_likely,1_good,Prakash and Sinha,2013,Deferred Revenue,CAR,continuous,Accounting,investment alt,2002,2007,DeferRev,1,EW,NA,NA,NA,"Annual change in deferred revenue (drc) scaled by average total assets (at) in t-1 and t. Exclude if negative book equity (ceq), deferred revenue equal to 0 in both years, revenue less than 5m, or SIC code between 6000 and 6999."
firm_age,firm_age.py,FirmAge,Predictor,2_likely,4_lack_data,Barry and Brown,1984,Firm age based on CRSP,JFE,continuous,Other,info proxy,1931,1980,FirmAge,-1,EW,NA,NA,exchcd==1,Months since start of CRSP coverage.
gr_lt_noa,gr_lt_noa.py,GrLTNOA,Predictor,2_likely,1_good,"Fairfield, Whisenant and Yohn",2003,Growth in long term operating assets,AR,continuous,Accounting,investment,1964,1993,LTNOAgr,1,EW,0.1,NA,abs(prc)>5,"Annual growth in net operating assets, minus accruals. Net operating assets are (rect + invt + ppent + aco + intan + ao- ap- lco- lo) / at. Accruals are ( rect-l12.rect + invt - l12.invt + aco - l12.aco - (ap - l12.ap + lco - l12.lco) - dp ) / ((at + l12.at)/2)"
gr_sale_to_gr_inv,gr_sale_to_gr_inv.py,GrSaleToGrInv,Predictor,2_likely,1_good,Abarbanell and Bushee,1998,Sales growth over inventory growth,AR,continuous,Accounting,sales growth,1974,1988,RevG2InvG,1,EW,NA,NA,NA,"Percentage growth in sales (sale) relative to average sales of t-1 and t-2, minus percentage growth in inventory (invt) relative to average inventory of t-1 and t-2. Both growth terms are calculated relative to t-1 only if t-2 is missing."
gr_sale_to_gr_overhead,gr_sale_to_gr_overhead.py,GrSaleToGrOverhead,Predictor,2_likely,1_good,Abarbanell and Bushee,1998,Sales growth over overhead growth,AR,continuous,Accounting,sales growth,1974,1988,RevG2OHG,1,EW,NA,NA,NA,"GrSaleToGrOverHead = Percentage growth in sales (sale) relative to average sales of t-1 and t-2, minus percentage growth in administrative expenses (xsga) relative to average administrative expenses of t-1 and t-2. Both growth terms are calculated relative to t-1 only if t-2 is missing."
herf_asset,herf_asset.py,HerfAsset,Predictor,2_likely,1_good,Hou and Robinson,2006,Industry concentration (assets),JF,continuous,Other,other,1963,2001,HerfAsset,-1,EW,0.2,NA,NA,"Three-year rolling average of the three digit industry Herfindahl index based on firm assets (at). Exclude regulated industries (4011, 4210, 4213 & year $\leq$ 1980; 4512 & year $\leq$ 1978, 4812, 4813 & year $\leq$ 1982, 4900-4999 in any year)"
high52,high52.py,High52,Predictor,2_likely,1_good,George and Hwang,2004,52 week high,JF,continuous,Price,momentum,1963,2001,High52,1,EW,0.3,NA,NA,Price (prc/cfacshr) divided by the maximum price over the previous 12 months.
mom_off_season_11yr_plus,mom_off_season_11yr_plus.py,MomOffSeason11YrPlus,Predictor,2_likely,1_good,Heston and Sadka,2008,Off season reversal years 11 to 15,JFE,continuous,Price,other,1965,2002,MomOffSeason11YrPlus,-1,EW,0.1,NA,"exchcd%in%c(1,2)",Average return in other months over the preceding 11-15 years.
mom_reversal,mom_reversal.py,MRreversal,Predictor,2_likely,3_distant,De Bondt and Thaler,1985,Medium-run reversal,JF,continuous,Price,long term reversal,1933,1980,Mom1813,-1,EW,0.2,NA,NA,Stock return between months t-18 and t-13.
oper_prof,oper_prof.py,OperProf,Predictor,2_likely,1_good,Fama and French,2006,operating profits / book equity,JFE,continuous,Accounting,profitability,1977,2003,ProfOper,1,EW,NA,NA,NA,"Revenue (revt) minus cost (cogs) - administrative expenses (xsga) - interest expenses (xint), scaled by book value of equity (ceq). Exclude smallest size tercile."
oper_prof_rd,oper_prof_rd.py,OperProfRD,Predictor,2_likely,1_good,Ball et al.,2016,Operating profitability R&D adjusted,JFE,continuous,Accounting,profitability,1963,2014,OperProfRD,1,VW,0.1,NYSE,NA,"Revenue (revt) minus cost (cogs) - (administrative expenses (xsga) - R&D expenses (xrd)), all divided by total assets (at) in year t. Replace all variables in the numerator with 0 if they are missing. Exclude if share code is greater 11, market value of equity, BM or total assets are missing, or if SIC code between 6000 and 6999."
real_estate,real_estate.py,realestate,Predictor,2_likely,1_good,Tuzel,2010,Real estate holdings,RFS,continuous,Accounting,asset composition,1971,2005,RealEstate,1,VW,0.2,NA,NA,"Industry-adjusted value of real estate holdings. Real estate holdings are (fatb+fatl)/ppegt if available, and (ppenb+ppenl)/ppent otherwise.  Drop firms in 2 digit sics with < 5 firm or missing at or missing both ppent and ppegt. Subtract monthly industry-mean at the 2 digit SIC level."
roe,roe.py,RoE,Predictor,2_likely,2_fair,Haugen and Baker,1996,net income / book equity,JFE,continuous,Accounting,profitability,1979,1993,RoE,1,EW,0.2,NA,abs(prc)>5,Net income (ni) over book value of equity (ceq). Exclude if price less than 5.
sp,sp.py,SP,Predictor,2_likely,1_good,"Barbee, Mukherji and Raines",1996,Sales-to-price,FAJ,continuous,Accounting,valuation,1979,1991,Rev2Price,1,EW,NA,NA,NA,Ratio of annual sales (sale) to market value of equity.
var_cf,var_cf.py,VarCF,Predictor,2_likely,2_fair,Haugen and Baker,1996,Cash-flow  to price variance,JFE,continuous,Accounting,cash flow risk,1979,1993,CF2Pvar,-1,EW,0.2,NA,NA,Rolling variance of (ib+dp)/mve\_c over the past 60 months (minimum 24 months data required).
vol_mkt,vol_mkt.py,VolMkt,Predictor,2_likely,2_fair,Haugen and Baker,1996,Volume to market equity,JFE,continuous,Trading,volume,1979,1993,Volume2Mkt,-1,EW,0.2,NA,abs(prc)>5,"Average monthly dollar trading volume (vol*abs(prc)) over the previous 12 months, scaled by market value of equity. Exclude if price less than 5."
vol_trend,vol_trend.py,VolumeTrend,Predictor,2_likely,2_fair,Haugen and Baker,1996,Volume Trend,JFE,continuous,Trading,volume,1979,1993,VolumeTrend,-1,EW,0.2,NA,NA,Rolling coefficient from regressing monthly trading volume on a linear time trend over a window of 60 months (require that at least 30 exist). Scale coefficient by 60-month average of trading volume.