notebook.community

Edit and run

Comparing Lazy Portfolios

This Quantopian Research based notebook compares SPY to the following portfolios:

  1. SPY 100%
  2. VTI/AGG 50/50 (stocks and bonds 50% / 50%)
  3. VTI/AGG 60/40
  4. VTI/AGG 70/30
  5. The Swensen Portfolio: TIP 15%, TLT 15%, VNQ 15%, EEM 10%, EFA 15%, VTI 30%
  6. My variation (POP): EDV 30%,VNQ 15%,VWO','EFA','VBR','MDY','RSP') context.pcts = [ 0.30, 0.15, 0.1, 0.15, 0.10, 0.0, 0.2] 




In [1]:



    


import pandas as pd
import matplotlib.pyplot as pyplot
import seaborn as sns



    











In [8]:



    


def stats(backtest):
    return backtest.risk.tail(1)



    











In [6]:



    


swensen = get_backtest('54fa0196e9787f0f115152aa')
pop = get_backtest('54fa094ca1094d0f12b31520')
spy = get_backtest('54fa0435f3bb8b0f0e66bb4f')



    


















    






100% Time: 0:00:08|###########################################################|
100% Time: 0:00:08|###########################################################|
100% Time: 0:00:05|###########################################################|

















In [10]:



    


pop = get_backtest('54fa094ca1094d0f12b31520')



    


















    






100% Time: 0:00:08|###########################################################|

















In [11]:



    


#swensen.risk.treasury_period_return.plot(label="Tresuries")
swensen.risk.benchmark_period_return.plot(label='SPY')
swensen.risk.period_return.plot(label="Swensen")
pop.risk.period_return.plot(label="POP")
pyplot.title("Portfolio Returns")
pyplot.legend(loc='best')
pyplot.ylabel("Portoflio % return")
pyplot.xlabel("Time")



    


















    
Out[11]:








<matplotlib.text.Text at 0x7fa56742c750>










    
























In [12]:



    


df_stats = stats(spy)
df_stats['Portfolio'] = 'SPY'
sw_stats = stats(swensen)
sw_stats['Portfolio'] = 'Swensen'
df_stats = df_stats.append(sw_stats)
pop_stats = stats(pop)
pop_stats['Portfolio'] = 'POP'
df_stats = df_stats.append(pop_stats)
df_stats = df_stats.set_index('Portfolio')
df_stats.transpose()



    


















    
Out[12]:










  
    

      Portfolio
      SPY
      Swensen
      POP
    

  
  
    

      alpha
      -0.00418823
       0.00554064
       0.04966755
    

    

      benchmark_period_return
        0.6348488
        0.6348488
        0.6348488
    

    

      benchmark_volatility
        0.2238107
        0.2238107
        0.2238107
    

    

      beta
        0.9778877
        0.7508968
        0.6977093
    

    

      excess_return
        0.5729296
        0.5329673
        0.8228448
    

    

      information
      -0.02595153
      -0.06333204
        0.1599096
    

    

      max_drawdown
        0.5264934
        0.4221693
        0.4020085
    

    

      period_label
          2015-02
          2015-02
          2015-02
    

    

      period_return
        0.5942296
        0.5542673
        0.8441448
    

    

      sharpe
        0.2823406
        0.3159118
        0.5285849
    

    

      sortino
        0.3578082
        0.4063945
        0.6968573
    

    

      treasury_period_return
           0.0213
           0.0213
           0.0213
    

    

      volatility
        0.2188839
        0.1779335
        0.1830342
    

  




















In [11]:



    


# this no longer works: swensen.benchmark_security.symbol



    











In [350]:



    


((result.risk.benchmark_period_return + 1) * result.capital_base).plot(label=result.benchmark_security.symbol)
((result.risk.period_return + 1) * result.capital_base).plot(label="Portfolio")
pyplot.title("Portfolio Returns")
pyplot.legend(loc='best')
pyplot.ylabel("Portoflio % return")
pyplot.xlabel("Time")



    


















    
Out[350]:








<matplotlib.text.Text at 0x7f1e6553fa10>










    
























In [328]:



    


benchmark_daily_close = get_pricing(swensen.benchmark_security.symbol, fields='close_price', 
                                    start_date= swensen.start_date,
                                    end_date = swensen.end_date,
                                    frequency='daily')

base_mult = swensen.capital_base / benchmark_daily_close[0]
benchmark_ending_portfolio_value = benchmark_daily_close * base_mult
benchmark_ending_portfolio_value[:5]



    


















    
Out[328]:








2007-03-02 00:00:00+00:00    100000.000000
2007-03-05 00:00:00+00:00     99071.808893
2007-03-06 00:00:00+00:00    100590.012952
2007-03-07 00:00:00+00:00    100345.373435
2007-03-08 00:00:00+00:00    101122.463664
Freq: C, Name: Security(8554 [SPY]), dtype: float64

















In [329]:



    


base_mult * 138



    


















    
Out[329]:








99294.862570153986

















In [330]:



    


benchmark_daily_close[:5]



    


















    
Out[330]:








2007-03-02 00:00:00+00:00    138.98
2007-03-05 00:00:00+00:00    137.69
2007-03-06 00:00:00+00:00    139.80
2007-03-07 00:00:00+00:00    139.46
2007-03-08 00:00:00+00:00    140.54
Freq: C, Name: Security(8554 [SPY]), dtype: float64

















In [ ]:



    


100



    











In [37]:



    


spy.cumulative_performance.ending_portfolio_value.plot(label="SPY",colormap=None)
swensen.cumulative_performance.ending_portfolio_value.plot(label="Swensen")
#benchmark_ending_portfolio_value.plot(label=swensen.benchmark_security.symbol)
pyplot.title("Portfolio Returns")
pyplot.legend(loc='best')
pyplot.ylabel("Portoflio Value in $")
pyplot.xlabel("Time")



    


















    
Out[37]:








<matplotlib.text.Text at 0x7f188a8a29d0>










    
























In [175]:



    


swensen.cumulative_performance.ending_portfolio_value[:5]



    


















    
Out[175]:








2007-03-02 00:00:00+00:00     99216.726551
2007-03-05 00:00:00+00:00     97801.685739
2007-03-06 00:00:00+00:00     99567.124901
2007-03-07 00:00:00+00:00     99273.012439
2007-03-08 00:00:00+00:00    100110.389799
Name: ending_portfolio_value, dtype: float64

















In [176]:



    


spy.cumulative_performance.ending_portfolio_value[:5]



    


















    
Out[176]:








2007-03-02 00:00:00+00:00     99164.60486
2007-03-05 00:00:00+00:00     98243.54486
2007-03-06 00:00:00+00:00     99750.08486
2007-03-07 00:00:00+00:00     99507.32486
2007-03-08 00:00:00+00:00    100278.44486
Name: ending_portfolio_value, dtype: float64

















In [197]:



    


# days with rebalancing transactions
transactions = [str(t.date()) for t in list(swensen.transactions.index.unique())]
transactions



    


















    
Out[197]:








['2007-03-02',
 '2008-01-09',
 '2008-10-09',
 '2008-11-21',
 '2009-07-27',
 '2010-05-06',
 '2010-05-26',
 '2012-05-08',
 '2013-11-13']

















In [178]:



    


# how many transactions ?
len(transactions)



    


















    
Out[178]:








9

















In [ ]:



    


stats(spy)



    











In [ ]:



    






    











In [297]:



    


#csv_VTHRX = fetch_url('http://ichart.yahoo.com/table.csv?s=VTHRX&a=0&b=0&c=2007&d=1&e=6&f=2015&g=d&ignore=.csv')
df_VTHRX = local_csv('VTHRX.csv')
df_VTHRX[:3]



    


















    
Out[297]:










  
    

      
      Date
      Open
      High
      Low
      Close
      Volume
      Adj Close
    

  
  
    

      0
       2015-02-06
       29.30
       29.30
       29.30
       29.30
       0
       29.30
    

    

      1
       2015-02-05
       29.47
       29.47
       29.47
       29.47
       0
       29.47
    

    

      2
       2015-02-04
       29.24
       29.24
       29.24
       29.24
       0
       29.24
    

  




















In [298]:



    


df_VTHRX.set_index('Date',inplace=True)
df_VTHRX.index = pd.to_datetime(df_VTHRX.index) # pd.DatetimeIndex(df_VTHRX.Date) # df_VTHRX.set_index('Date')
df_VTHRX.sort(inplace=True)
df_VTHRX = df_VTHRX[pd.datetime(2007, 3, 2):]
VTHRX_base_mult = swensen.capital_base / df_VTHRX['Adj Close'][0]
df_VTHRX['ending_portfolio_value'] = df_VTHRX['Adj Close'] * VTHRX_base_mult
df_VTHRX[:5]



    


















    
Out[298]:










  
    

      
      Open
      High
      Low
      Close
      Volume
      Adj Close
      ending_portfolio_value
    

  
  
    

      2007-03-02
       22.34
       22.34
       22.34
       22.34
       0
       18.88
       100000.000000
    

    

      2007-03-05
       22.09
       22.09
       22.09
       22.09
       0
       18.67
        98887.711864
    

    

      2007-03-06
       22.44
       22.44
       22.44
       22.44
       0
       18.97
       100476.694915
    

    

      2007-03-07
       22.41
       22.41
       22.41
       22.41
       0
       18.94
       100317.796610
    

    

      2007-03-08
       22.58
       22.58
       22.58
       22.58
       0
       19.09
       101112.288136
    

  




















In [306]:



    


df_VTHRX.ending_portfolio_value.plot(label="VTHRX")
swensen.cumulative_performance.ending_portfolio_value.plot(label="Swensen")
#spy.cumulative_performance.ending_portfolio_value.plot(label="SPY", style=[':'])
pyplot.title("Portfolio Returns")
pyplot.legend(loc='best')
pyplot.ylabel("Portoflio Value in $")
pyplot.xlabel("Time")



    


















    
Out[306]:








<matplotlib.text.Text at 0x7f1e68d94410>










    
























In [72]:



    


M5050 = get_backtest('54ea9bd57c6f390f1639a55f')
M6040 = get_backtest('54ea9beba77a8b0f24e91f8c')
M7030 = get_backtest('54ea9bff52d9f50f07c03985')



    


















    






100% Time: 0:00:07|###########################################################|
100% Time: 0:00:06|###########################################################|
100% Time: 0:00:07|###########################################################|

















In [48]:



    


M5050.risk.period_return.plot(label="50/50")
M6040.risk.period_return.plot(label="60/40")
M7030.risk.period_return.plot(label="70/30")
M5050.risk.benchmark_period_return.plot(label='SPY')
pyplot.title("VTI / AGG Portfolio Returns")
pyplot.legend(loc='best')
pyplot.ylabel("Portoflio % return")
pyplot.xlabel("Time")



    


















    
Out[48]:








<matplotlib.text.Text at 0x7f188a8d0190>










    
























In [74]:



    


df_stats = stats(spy)
df_stats['Portfolio'] = 'SPY'
M5050_stats = stats(M5050)
M5050_stats['Portfolio'] = '50/50'
df_stats = df_stats.append(M5050_stats)
M6040_stats = stats(M6040)
M6040_stats['Portfolio'] = '60/40'
df_stats = df_stats.append(M6040_stats)
M7030_stats = stats(M7030)
M7030_stats['Portfolio'] = '70/30'
df_stats = df_stats.append(M7030_stats)
df_stats = df_stats.set_index('Portfolio')
df_stats.transpose()



    


















    
Out[74]:










  
    

      Portfolio
      SPY
      50/50
      60/40
      70/30
    

  
  
    

      alpha
        0.001307798
       0.02107049
       0.02066013
       0.01940546
    

    

      benchmark_period_return
          0.7411849
        0.7411849
        0.7411849
        0.7411849
    

    

      benchmark_volatility
          0.2168138
        0.2168138
        0.2168138
        0.2168138
    

    

      beta
          0.9827479
        0.4520165
        0.5532459
         0.656437
    

    

      excess_return
          0.7207527
        0.5806055
        0.6347158
        0.6830577
    

    

      information
       0.0004995468
       -0.1654271
      -0.08460003
       -0.0311927
    

    

      max_drawdown
           0.541778
        0.2972374
        0.3520304
        0.4043448
    

    

      period_label
            2015-02
          2015-02
          2015-02
          2015-02
    

    

      period_return
          0.7420527
        0.6019055
        0.6560158
        0.7043577
    

    

      sharpe
          0.3272416
        0.5086997
        0.4791808
         0.449542
    

    

      sortino
           0.409983
        0.6549773
        0.6157634
        0.5744546
    

    

      treasury_period_return
             0.0213
           0.0213
           0.0213
           0.0213
    

    

      volatility
          0.2131164
        0.1032952
        0.1235127
        0.1448493
    

  




















In [ ]:

Content source: paulperry/quant

Similar notebooks:

notebook.community | gallery | about