The effect of Fed rate hikes on Indian indicators

# Load the packages:
using TSFrames
using MarketData
using Dates
using DataFrames
using EventStudies
using CairoMakie

First, get data about stock prices from MarketData.jl:

NIFTY index (India)

nifty = TSFrame(MarketData.yahoo("^NSEI"))
select!(nifty.coredata, :Index, :AdjClose => :NIFTY);

3838×1 TSFrame with Dates.Date Index
 Index       NIFTY
 Date        Float64?
──────────────────────
 2007-09-17   4494.65
 2007-09-18   4546.2
 2007-09-19   4732.35
 2007-09-20   4747.55
 2007-09-21   4837.55
 2007-09-24   4932.2
 2007-09-25   4938.85
 2007-09-26   4940.5
     ⋮          ⋮
 2023-03-21  17107.5
 2023-03-22  17151.9
 2023-03-23  17076.9
 2023-03-24  16945.1
 2023-03-27  16985.7
 2023-03-28  16951.7
 2023-03-29  17080.7
     3823 rows omitted

USD/INR exchange rate

usd_inr = TSFrame(MarketData.yahoo("USDINR=X"))
select!(usd_inr.coredata, :Index, :AdjClose => :USDINR);

5045×1 TSFrame with Dates.Date Index
 Index       USDINR
 Date        Float64?
──────────────────────
 2003-12-01   45.71
 2003-12-02   45.629
 2003-12-03   45.55
 2003-12-04   45.548
 2003-12-05   45.449
 2003-12-08   45.47
 2003-12-09   45.431
 2003-12-10   45.371
     ⋮          ⋮
 2023-03-23   82.5258
 2023-03-24   82.2653
 2023-03-27   82.3388
 2023-03-28   82.2149
 2023-03-29   82.1643
 2023-03-30   82.2554
 2023-03-31   82.09
     5030 rows omitted

US Fed funds rate

fed_rate = TSFrame(MarketData.fred("FEDFUNDS"))
fed_rate_2007 = TSFrame(fed_rate.coredata[searchsortedfirst(index(fed_rate), Date(2007, 1, 1)):end, :])
fed_rate_diffs = diff(fed_rate_2007)
dropmissing!(fed_rate_diffs.coredata)
fed_rate_dates = filter(:VALUE => >(0), fed_rate_diffs.coredata)

83×2 DataFrame

Row	Index	VALUE
	Date	Float64
1	2007-02-01	0.01
2	2007-07-01	0.01
3	2008-06-01	0.02
4	2008-07-01	0.01
5	2009-02-01	0.07
6	2009-05-01	0.03
7	2009-06-01	0.03
8	2010-02-01	0.02
9	2010-03-01	0.03
10	2010-04-01	0.04
11	2010-08-01	0.01
12	2011-08-01	0.03
13	2011-11-01	0.01
14	2012-01-01	0.01
15	2012-02-01	0.02
16	2012-03-01	0.03
17	2012-04-01	0.01
18	2012-05-01	0.02
19	2012-09-01	0.01
20	2012-10-01	0.02
21	2013-02-01	0.01
22	2013-04-01	0.01
23	2013-10-01	0.01
24	2013-12-01	0.01
25	2014-03-01	0.01
26	2014-04-01	0.01
27	2014-06-01	0.01
28	2014-12-01	0.03
29	2015-04-01	0.01
30	2015-06-01	0.01
31	2015-08-01	0.01
32	2015-12-01	0.12
33	2016-01-01	0.1
34	2016-02-01	0.04
35	2016-04-01	0.01
36	2016-06-01	0.01
37	2016-07-01	0.01
38	2016-08-01	0.01
39	2016-11-01	0.01
40	2016-12-01	0.13
41	2017-01-01	0.11
42	2017-02-01	0.01
43	2017-03-01	0.13
44	2017-04-01	0.11
45	2017-05-01	0.01
46	2017-06-01	0.13
47	2017-07-01	0.11
48	2017-08-01	0.01
49	2017-11-01	0.01
50	2017-12-01	0.14
51	2018-01-01	0.11
52	2018-02-01	0.01
53	2018-03-01	0.09
54	2018-04-01	0.18
55	2018-05-01	0.01
56	2018-06-01	0.12
57	2018-07-01	0.09
58	2018-09-01	0.04
59	2018-10-01	0.24
60	2018-11-01	0.01
61	2018-12-01	0.07
62	2019-01-01	0.13
63	2019-03-01	0.01
64	2019-04-01	0.01
65	2019-07-01	0.02
66	2020-02-01	0.03
67	2020-06-01	0.03
68	2020-07-01	0.01
69	2020-08-01	0.01
70	2021-06-01	0.02
71	2021-07-01	0.02
72	2022-03-01	0.12
73	2022-04-01	0.13
74	2022-05-01	0.44
75	2022-06-01	0.44
76	2022-07-01	0.47
77	2022-08-01	0.65
78	2022-09-01	0.23
79	2022-10-01	0.52
80	2022-11-01	0.7
81	2022-12-01	0.32
82	2023-01-01	0.23
83	2023-02-01	0.24

Then, we perform the event studies!

NIFTY

eventtime_returns_ts, event_status = EventStudies.eventstudy(
    levels_to_returns(nifty),
    (:NIFTY,) .=> fed_rate_dates.Index,
    -6:7
    )

eventtime_cumulative_ts = remap_cumsum(eventtime_returns_ts)
t0, lower, upper = inference(BootstrapInference(), eventtime_cumulative_ts)

eventtime_cumulative_ts

14×1 TSFrame with Int64 Index
 Index  NIFTY
 Int64  Float64?
────────────────────
    -6   0.0
    -5  -0.00137981
    -4  -1.25855
    -3  -2.87903
    -2  -2.62601
    -1  -2.55124
     0  -2.81117
     1  -2.84467
     2  -1.47059
     3  -1.97286
     4  -2.21577
     5  -1.37179
     6  -1.25016
     7  -1.45687

plot the result

f, a, p = scatterlines(index(eventtime_cumulative_ts), t0; label = "Mean over all events")
p2 = band!(a, index(eventtime_cumulative_ts), lower, upper; label = "95% CI")
translate!(p, 0, 0, 1) # bring the scatterlines to the front

leg = axislegend(a; position = :rt)

a.title = "NIFTY index performance after a Fed rate hike"
a.subtitle = rich(rich("t=0", font = to_font("Fira Mono")), " indicates the time of rate hike")
a.titlealign = :left
a.xlabel = "Event time (days)"
a.ylabel = "NIFTY returns (%)"
a.xticks = WilkinsonTicks(8; k_min = 5, k_max = 10)
f

save("nifty.pdf", f; ptperunit = 1)

USD/INR

eventtime_returns_ts, event_status = EventStudies.eventstudy(levels_to_returns(usd_inr), (:USDINR,) .=> fed_rate_dates.Index, -6:7)

eventtime_cumulative_ts = remap_cumsum(eventtime_returns_ts)
t0, lower, upper = inference(BootstrapInference(), eventtime_cumulative_ts)

eventtime_cumulative_ts

14×1 TSFrame with Int64 Index
 Index  USDINR
 Int64  Float64?
───────────────────
    -6   0.0
    -5   0.139676
    -4   0.0180403
    -3  -0.0526597
    -2   0.0245454
    -1   0.112719
     0   0.343513
     1   0.272427
     2   0.682306
     3   0.901577
     4   1.54286
     5   1.52703
     6   1.4274
     7   1.27449

plot the result

f, a, p = scatterlines(index(eventtime_cumulative_ts), t0; label = "Mean over all events")
p2 = band!(a, index(eventtime_cumulative_ts), lower, upper; label = "95% CI")
translate!(p, 0, 0, 1) # bring the scatterlines to the front

leg = axislegend(a; position = :rt)

a.title = "USD-INR performance after a Fed rate hike"
a.subtitle = rich(rich("t=0", font = to_font("Fira Mono")), " indicates the time of rate hike")
a.titlealign = :left
a.xlabel = "Event time (days)"
a.ylabel = "USD-INR returns (%)"
a.xticks = WilkinsonTicks(8; k_min = 5, k_max = 10)
f

save("usdinr2.pdf", f; ptper_unit = 1)

This page was generated using Literate.jl.