misc-research/Financial-Marginalization-and-Cryptocurrency/Financial-Marginalization-and-Cryptocurrency-Payments.Rmd

---
title: "Financial Marginalization and Cryptocurrency Payments"
author: "Rucknium"
date: '2022-05-26'
output: html_document
---

```{css, echo=FALSE}
body .main-container {
  max-width: 1480px !important;
  width: 1480px !important;
}
body {
  max-width: 1480px !important;
}
```



```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE, size = "small", tidy = FALSE)
options(width = 170)



# or_glm function adapted from "oddsratio" package so it can
# handle svyglm functions:

or_svyglm <- function (data, model, incr, ci = 0.95) 
{
  if (any(class(model) %in% "glm")) {
    preds <- names(coefficients(model))[2:length(coefficients(model))]
    coef <- coefficients(model)[2:length(coefficients(model))]
  }
  increments <- list()
  odds_ratios <- list()
  ci_low <- list()
  ci_high <- list()
  for (i in preds) {
    if (any(class(model) %in% "glm")) {
      ci_list <- data.frame(suppressMessages(confint(model, 
        level = ci)))[-1, ]
    }
    if (is.numeric(data[[i]]) | is.integer(data[[i]])) {
      odds_ratios[[i]] <- round(exp(as.numeric(coef[[i]]) * 
          as.numeric(incr[[i]])), 3)
      if (!class(model)[1] == "glmmPQL") {
        ci_low[[i]] <- round(exp(ci_list[i, 1] * as.numeric(incr[[i]])), 
          3)
        ci_high[[i]] <- round(exp(ci_list[i, 2] * as.numeric(incr[[i]])), 
          3)
      }
      increments[[i]] <- as.numeric(incr[[i]])
      or <- odds_ratios[[i]]
    }
    else {
      odds_ratios[[i]] <- round(exp(as.numeric(coef[[i]])), 
        3)
      if (!class(model)[1] == "glmmPQL") {
        ci_low[[i]] <- round(exp(ci_list[i, 1]), 3)
        ci_high[[i]] <- round(exp(ci_list[i, 2]), 3)
      }
      increments[[i]] <- "Indicator variable"
      or <- odds_ratios[[i]]
    }
  }
  result <- data.frame(predictor = names(odds_ratios), oddsratio = unlist(odds_ratios, 
    use.names = FALSE), ci_low = unlist(ci_low, use.names = FALSE), 
    ci_high = unlist(ci_high, use.names = FALSE), increment = unlist(increments, 
      use.names = FALSE))
  if (any(class(model) %in% "glm")) {
    col_names <- gsub("\\.\\.", replacement = "", names(ci_list))
    col_names <- gsub("X", replacement = "", col_names)
    colnames(result)[3] <- paste0("ci_low (", col_names[1], 
      ")")
    colnames(result)[4] <- paste0("ci_high (", col_names[2], 
      ")")
  }
  return(result)
}

```

## Load packages

```{r load-packages}
# install.packages("readstata13")
# install.packages("survey")
# install.packages("corrplot")

library(readstata13)
library(survey)
library(corrplot)

```


## Download and import data

```{r get-data}

# download.file("https://www.federalreserve.gov/consumerscommunities/files/SHED_public_use_data_2021_(Stata).zip", 
#   "data/SHED_public_use_data_2021_(Stata).zip")
# unzip("data/SHED_public_use_data_2021_(Stata).zip", exdir = "data")


SHED <- readstata13::read.dta13("data/public2021.dta", generate.factors = TRUE, nonint.factors = TRUE)

# S16_b
# In the past year, have you done the following
# with cryptocurrency, such as Bitcoin or Ethereum?
# - Used to buy something or make a payment
SHED$crypto.used.as.payment <- SHED$S16_b == "Yes"

# BK1
# Do you and/or your spouse or partner
# currently have a checking, savings or money
# market account?
SHED$lacks.bank.account <- SHED$BK1 == "No"

# C2A
# Do you currently have at least one credit card?
SHED$lacks.credit.card <- SHED$C2A == "No"

# ppgender
# Gender [Ipsos source]
SHED$is.male <- SHED$ppgender == "Male"

# race_5cat Race 5 categories
SHED$race <- SHED$race_5cat

# ED0
# What is the highest level of school you have
# completed or the highest degree you have
# received?
SHED$education.level <- relevel(SHED$ED0, "High school degree or GED")

# ED1
# Which one of the following broad categories
# best describes your (current/most recent)
# educational program?
SHED$education.subject <- relevel(SHED$ED1, "Business/management")

# pppa_lgb [Ipsos source]
# Q230: Which of the following best describes how you think of yourself?
SHED$lgbtq <- relevel(SHED$pppa_lgb, "Straight, that is, not gay")

# ppp20197 [Ipsos source]
# QEG22: Are you a citizen of the United States?
SHED$is.noncitizen <- SHED$ppp20197 == "No"

# I40
# Which of the following categories best
# describes the total income that you and/or
# your spouse or partner received from all
# sources, before taxes and deductions, in the
# past 12 months?
SHED$income.category <- SHED$I40

# B2
# Overall, which one of the following best
# describes how well you are managing
# financially these days?
SHED$overall.financial.wellbeing <- SHED$B2

# I41_b
# Supplemental Nutrition Assistance Program
# (SNAP or food stamps) - In the past 12
# months, have you received any of the
# following?
SHED$received.food.stamps <- SHED$I41_b

# FL0
# On a scale from zero to ten, where zero is
# not at all willing to take risks and ten is
# very willing to take risks, what number would
# you be on the scale?

SHED$risk.tolerance <- as.numeric(SHED$FL0) - 1

# B10
# Overall, on a scale from zero to ten, where
# zero is not at all satisfied and ten is
# completely satisfied, how satisfied are you
# with life as a whole these days?
SHED$life.satisfaction <- as.numeric(SHED$B10) - 1

# ppage
# Age [Ipsos source]
# ppcmdate
# Date member completed Core survey
# Must correct age variable for time of initial Ipsos survey
SHED$age <- SHED$ppage + (2021 - as.numeric(substr(SHED$ppcmdate, 1, 4)))

# ind1
# IND1: Industry (tight scale) in current or main job
SHED$job.industry <- relevel(SHED$ind1, "Retail/Stores/Shopping (including Online Retail)")

# ppcm0160 [Ipsos source]
# Q26: Occupation (detailed) in current or main job
SHED$job.occupation <- relevel(SHED$ppcm0160, "Retail Sales")

# ppcm1301 [Ipsos source]
# GOVEMP1: Employer type
SHED$employer.type <- relevel(SHED$ppcm1301, "Private-for-profit company")

# ppmsacat
# MSA Status [Ipsos source]
SHED$resides.in.metro.statistical.area <- SHED$ppmsacat == "Metro"

# ppfs0596 [Ipsos source]
# Q22: What is the approximate total amount of
# your household's savings and investments?
SHED$total.household.savings <- relevel(SHED$ppfs0596, "$100,000 - $249,999")

# A1_a
# In the past 12 months, has each of the following happened to you:
# - Turned down for credit
SHED$rejected.for.credit <- SHED$A1_a == "Yes"

# BK2_a
# In the past 12 months, did you and/or spouse or partner:
# - Purchase a money order from a place other than a bank
SHED$purchase.non.bank.money.order <- SHED$BK2_a == "Yes"

# BK2_b
# In the past 12 months, did you and/or spouse or partner:
# - Cash a check at a place other than a bank
SHED$cash.check.non.bank <- SHED$BK2_b == "Yes"

# BK2_c
# In the past 12 months, did you and/or spouse or partner:
# - Take out a payday loan or payday advance
SHED$take.payday.loan <- SHED$BK2_c == "Yes"

# BK2_d
# In the past 12 months, did you and/or spouse or partner:
# - Take out a pawn shop loan or an auto title loan
SHED$take.auto.or.pawn.shop.loan <- SHED$BK2_d == "Yes"

# BK2_e
# In the past 12 months, did you and/or spouse or partner:
# - Obtain a tax refund advance to receive your refund faster
SHED$take.tax.refund.advance <- SHED$BK2_e == "Yes"

# BNPL1
# In the past year, have you used a “Buy Now
# Pay Later” service to buy something?
SHED$used.buy.now.pay.later <- SHED$BNPL1 == "Yes"

# ppfs1482 [Ipsos source]
# Q108: Where do you think your credit score falls
SHED$perceived.credit.score <- relevel(SHED$ppfs1482, "Fair")

# GE2A
# Some people earn money by selling items at
# places like flea markets and garage sales or
# through online marketplaces like eBay or
# Etsy. In the past month, have you made money
# by selling items in any of these ways?

SHED$informal.selling.of.goods <- SHED$GE2A == "Yes"

# GE1A
# In the past month, have you done any
# freelance or gig-work, either to supplement
# your income or as your main job?

SHED$freelance.or.gig.work <- SHED$GE1A == "Yes"

# E7
# During the past 12 months, have you
# personally experienced discrimination or
# unfair treatment because of your race,
# ethnicity, age, religion, disability status,
# sexual orientation, gender, or gender
# identity?
SHED$experienced.discrimination <- SHED$E7 == "Yes"

# E8_b 
# In the past 12 months, did you personally experience
# discrimination or unfair treatment while
# doing any of the following?
# - Banking or applying for a loan
SHED$experienced.discrimination.in.banking <- SHED$E8_b %in% "Yes"

# xlaptop
# Is R a KP laptop user?
SHED$is.kp.laptop.user <- SHED$xlaptop %in% "Yes"

# devicetype2
# DOV: Device Type - at the end of survey
SHED$respondent.device.type <- relevel(SHED$devicetype2, "WinPC")


SHED <- svydesign(ids = ~0, data = SHED, weights = SHED$weight_pop)

# weight_pop used as survey weights, in accordance with suggestion by:
# https://www.federalreserve.gov/consumerscommunities/files/SHED_2021codebook.pdf


```

## Checking correlations between main variables


```{r initial-correlations}


svyvar.covariance <- svyvar(~ age + is.male + crypto.used.as.payment +
    lacks.bank.account + lacks.credit.card, SHED, na.rm = TRUE)
attr(svyvar.covariance, "var") <- NULL
svyvar.correlation <- cov2cor(as.matrix(svyvar.covariance))

print(round(svyvar.correlation, 3))

corrplot(svyvar.correlation, tl.col = "darkred", tl.srt = 35, 
  method = "shade", number.digits = 2, addshade = "all", diag = FALSE,
         title = "\n\n Correlation Matrix of Financial Marginalization\nand Use of Cryptocurrency as a Means of Payment",
         addCoef.col = "black", type = "lower")

```

# Main results

## lacks.bank.account

```{r main-results-lacks-bank-account}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + lacks.bank.account, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## lacks.credit.card

```{r main-results-lacks-credit-card}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + lacks.credit.card, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))


```

# Additional results on financial marginalization

## rejected.for.credit

```{r other-financial-marginalization-rejected-for-credit}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + rejected.for.credit, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## purchase.non.bank.money.order

```{r other-financial-marginalization-purchase-non-bank-money-order}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + purchase.non.bank.money.order, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## cash.check.non.bank

```{r other-financial-marginalization-cash-check-non-bank}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + cash.check.non.bank, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## take.payday.loan

```{r other-financial-marginalization-take-payday-loan}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + take.payday.loan, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## take.auto.or.pawn.shop.loan

```{r other-financial-marginalization-take-auto-or-pawn-shop-loan}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + take.auto.or.pawn.shop.loan, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## take.tax.refund.advance

```{r other-financial-marginalization-take-tax-refund-advance}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + take.tax.refund.advance, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## used.buy.now.pay.later

```{r other-financial-marginalization-used-buy-now-pay-later}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + used.buy.now.pay.later, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## experienced.discrimination.in.banking

```{r other-financial-marginalization-experienced-discrimination-in-banking}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + experienced.discrimination.in.banking, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## perceived.credit.score

```{r other-financial-marginalization-perceived-credit-score}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + perceived.credit.score, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

# General societal marginalization

## experienced.discrimination

```{r general-societal-marginalization-experienced-discrimination}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + experienced.discrimination, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## race

```{r general-societal-marginalization-race}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + race, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## is.noncitizen

```{r general-societal-marginalization-is.noncitizen}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + is.noncitizen, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## lgbtq

```{r general-societal-marginalization-lgbtq}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + lgbtq, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

# Data quality sanity checks

## education.subject

```{r data-quality-checks-education-subject}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + education.subject, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## job.industry

```{r data-quality-checks-job-industry}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + job.industry, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## job.occupation

```{r data-quality-checks-job-occupation}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + job.occupation, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## risk.tolerance

```{r data-quality-checks-risk-tolerance}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + risk.tolerance, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1, risk.tolerance = 1))


```


# Miscellaneous

## education.level

```{r miscellaneous-education-level}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + education.level, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## income.category

```{r miscellaneous-income-category}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + income.category, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## total.household.savings

```{r miscellaneous-total-household-savings}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + total.household.savings, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## overall.financial.wellbeing

```{r miscellaneous-overall-financial-wellbeing}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + overall.financial.wellbeing, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## received.food.stamps

```{r miscellaneous-received-food-stamps}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + received.food.stamps, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## life.satisfaction

```{r miscellaneous-life-satisfaction}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + life.satisfaction, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1, life.satisfaction = 1))

```

## employer.type

```{r miscellaneous-employer-type}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + employer.type, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## employer.type = Self-employed & job.occupation = Computer and Mathematical

```{r miscellaneous-self-employed-computer-industry}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male +
    I(employer.type == "Self-employed") * I(job.occupation == "Computer and Mathematical"), 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
print(or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1)))

```

## informal.selling.of.goods

```{r miscellaneous-informal-selling-of-goods}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + informal.selling.of.goods, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## freelance.or.gig.work

```{r miscellaneous-freelance-or-gig-work}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + freelance.or.gig.work, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## resides.in.metro.statistical.area

```{r miscellaneous-resides-in-metro-statistical-area}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + resides.in.metro.statistical.area, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## respondent.device.type

```{r miscellaneous-respondent-device-type}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + respondent.device.type, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))

```

## is.kp.laptop.user

```{r miscellaneous-is-kp-laptop-user}

svyglm.fit <- svyglm(crypto.used.as.payment ~ age + is.male + is.kp.laptop.user, 
  SHED, family = stats::quasibinomial(link = "logit"))
summary(svyglm.fit)
or_svyglm(SHED$variables, svyglm.fit, incr = list(age = 1))


```