G'day ibm1620,

"But I can't get over the feeling that there is a much simpler solution that's eluded me."

I knocked this off in my lunch break: I expect there are many opportunities for improvement; however, it's certainly simpler than the code you posted.

#!/usr/bin/env perl

use v5.36;
use autodie;
use constant {
    DATE => 0,
    AMOUNT => 1,
    CATEGORY => 2,
    DESCRIPTION => 3,
};

use Text::CSV;

my $csv_file = 'pm_11145274_tabulate_csv.csv';

# Dummy figures from 1st pass which you say you know how to do
my @max_widths = qw{10 6 15 55};

say 'Shrunk to fit 52-char-wide window';
tabulate($csv_file, \@max_widths, 52);
say 'Shrunk to fit 46-char-wide window';
tabulate($csv_file, \@max_widths, 46);

sub tabulate ($filename, $max_widths, $table_width) {
    state $des_weight = 0.7;
    my $table_available = $table_width - 4; # 4 pipe chars
    my $data_fmt = '%8d|'; # date: YYYYMMDD
    my $head_fmt = '%-8s|';
    $table_available -= 8;
    $data_fmt .= '%6.2f|'; # amount: | 12.34|
    $head_fmt .= '%-6s|';
    $table_available -= 6;
    my $max_cat_des = $max_widths->[CATEGORY] + $max_widths->[DESCRIPT
+ION];
    my $cat_width = int($table_available * ($max_widths->[CATEGORY] / 
+$max_cat_des));
    my $des_width = int($table_available * ($max_widths->[DESCRIPTION]
+ / $max_cat_des));
    $des_width = int($des_width * $des_weight);
    $cat_width += $table_available - ($cat_width + $des_width);
    $data_fmt .= "%-${cat_width}.${cat_width}s|%-${des_width}.${des_wi
+dth}s|\n";
    $head_fmt .= "%-${cat_width}s|%-${des_width}s|\n";

    my $csv = Text::CSV::->new();

    {
        open my $fh, '<', $filename;
        my $head_row = $csv->getline($fh);
        printf $head_fmt, @$head_row;

        while (my $data_row = $csv->getline($fh)) {
            printf $data_fmt, $data_row->[DATE] =~ y/-//dr,
                    $data_row->@[AMOUNT, CATEGORY, DESCRIPTION];
        }
    }

    return;
}
[download]

I used this for input. You may need some code adjustments depending on what your real input looks like.

$ cat pm_11145274_tabulate_csv.csv
Date,Amount,Category,Description
2022-06-23,123.45,Software & Tech,BACKBLAZE HTTPSWWW.BACKCA
2022-06-24,63.45,Internet,RECURRING PAYMENT AUTHORIZED ON 06/11 SPECTR
+UM TX
2022-06-24,69.34,Phone,RECURRING PAYMENT AUTHORIZED ON 06/02 VZWRLSS*A
+POCC VISE
[download]

I reduced the YYYY-MM-DD date to YYYYMMDD to free up a couple of extra characters and aligned the amounts. Fiddle with the arithmetic to get exactly what you want: adjusting $des_weight may be all that you need. Here's the output:

$ ./pm_11145274_tabulate_csv.pl
Shrunk to fit 52-char-wide window
Date    |Amount|Category        |Description       |
20220623|123.45|Software & Tech |BACKBLAZE HTTPSWWW|
20220624| 63.45|Internet        |RECURRING PAYMENT |
20220624| 69.34|Phone           |RECURRING PAYMENT |
Shrunk to fit 46-char-wide window
Date    |Amount|Category     |Description    |
20220623|123.45|Software & Te|BACKBLAZE HTTPS|
20220624| 63.45|Internet     |RECURRING PAYME|
20220624| 69.34|Phone        |RECURRING PAYME|
[download]