TST: Correct test to use trend not level

[bashtage]

Correct ARIMA(1,1,1) to have a non-zero trend to avoid random test failures

• tests added / passed.
• code/documentation is well formatted.
• properly formatted commit message. See
  NumPy's guide.

Notes:

• It is essential that you add a test when making code changes. Tests are not
  needed for doc changes.
• When adding a new function, test values should usually be verified in another package (e.g., R/SAS/Stata).
• When fixing a bug, you must add a test that would produce the bug in master and
  then show that it is fixed with the new code.
• New code additions must be well formatted. Changes should pass flake8. If on Linux or OSX, you can
  verify you changes are well formatted by running

  git diff upstream/master -u -- "*.py" | flake8 --diff --isolated
  

  assuming flake8 is installed. This command is also available on Windows
  using the Windows System for Linux once flake8 is installed in the
  local Linux environment. While passing this test is not required, it is good practice and it help
  improve code quality in statsmodels.
• Docstring additions must render correctly, including escapes and LaTeX.

[codecov]

Codecov Report

Merging #6020 into master will decrease coverage by <.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master    #6020      +/-   ##
==========================================
- Coverage   83.46%   83.46%   -0.01%     
==========================================
  Files         606      606              
  Lines       95638    95639       +1     
  Branches    10411    10411              
==========================================
  Hits        79829    79829              
- Misses      13436    13443       +7     
+ Partials     2373     2367       -6

Impacted Files	Coverage Δ
statsmodels/tsa/tests/test_arima.py	99.52% <100%> (ø)	⬆️
statsmodels/tools/grouputils.py	69.46% <0%> (-0.45%)	⬇️
statsmodels/sandbox/distributions/mv_normal.py	38.82% <0%> (ø)	⬆️
statsmodels/sandbox/regression/treewalkerclass.py	0% <0%> (ø)	⬆️
statsmodels/sandbox/stats/multicomp.py	46.93% <0%> (ø)	⬆️
statsmodels/stats/libqsturng/make_tbls.py	0% <0%> (ø)	⬆️
statsmodels/tsa/statespace/mlemodel.py	85.11% <0%> (ø)	⬆️
statsmodels/iolib/summary.py	58.15% <0%> (ø)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 58c52b1...05ff24a. Read the comment docs.

[coveralls]

Coverage increased (+0.03%) to 85.951% when pulling 05ff24a on bashtage:fix-buggy-test into 58c52b1 on statsmodels:master.